{"data":{"records":[{"id":"hiq-labs-inc-v-linkedin-corp-273-f-supp-3d-1099-","dataset":"scraping","jurisdiction":"US federal (Ninth Circuit / N.D. Cal.)","instrument":"hiQ Labs, Inc. v. LinkedIn Corp., 273 F. Supp. 3d 1099 (N.D. Cal. 2017) (preliminary injunction)","status":"in-force","summary":"Does the CFAA bar scraping publicly available website data? — At the preliminary-injunction stage the court found hiQ raised serious questions that LinkedIn could not invoke the CFAA to block scraping of PUBLIC profiles, because CFAA 'without authorization' presupposes access to information not otherwise open to the public. This was a likelihood-of-success / balance-of-hardships finding, NOT a final merits holding.","source_url":"https://docs.justia.com/cases/federal/district-courts/california/candce/3:2017cv03301/312704/199","source_status":"blocked","not_live_law":false,"detail":{"question":"Does the CFAA bar scraping publicly available website data?","holding":"At the preliminary-injunction stage the court found hiQ raised serious questions that LinkedIn could not invoke the CFAA to block scraping of PUBLIC profiles, because CFAA 'without authorization' presupposes access to information not otherwise open to the public. This was a likelihood-of-success / balance-of-hardships finding, NOT a final merits holding.","instrument":"hiQ Labs, Inc. v. LinkedIn Corp., 273 F. Supp. 3d 1099 (N.D. Cal. 2017) (preliminary injunction)","decision_date":"2017-08-14","authority":"U.S. District Court, N.D. Cal. (Judge Edward Chen)","status":"narrowed","binding_scope":"Not binding precedent; a district-court preliminary-injunction order in one case. Persuasive only.","applies_to":["public-data","technical-measures"],"practical_rule":"A CFAA claim is weak against scraping of genuinely public (no-login) pages, but a PI ruling is not a licence — other claims survive.","source_type":"court-opinion","notes":"Preliminary posture. The full case history matters: hiQ ultimately LOST (see contract rows below)."}},{"id":"hiq-labs-inc-v-linkedin-corp-938-f-3d-985-9th-ci","dataset":"scraping","jurisdiction":"US federal (Ninth Circuit)","instrument":"hiQ Labs, Inc. v. LinkedIn Corp., 938 F.3d 985 (9th Cir. 2019), vacated and remanded 141 S. Ct. 2752 (2021), adhered to on remand 31 F.4th 1180 (9th Cir. 2022)","status":"in-force","summary":"Does the CFAA 'without authorization' provision reach scraping of public data? — The Ninth Circuit AFFIRMED the preliminary injunction, holding hiQ raised serious questions that the CFAA does not apply where data is publicly available without authentication ('the concept of authorization presupposes access restrictions'). Reaffirmed on remand after the Supreme Court GVR in light of Van Buren. Still a serious-questions / PI standard, not a full merits adjudication of the CFAA on a complete record.","source_url":"https://cdn.ca9.uscourts.gov/datastore/opinions/2022/04/18/17-16783.pdf","source_status":"verified","not_live_law":false,"detail":{"question":"Does the CFAA 'without authorization' provision reach scraping of public data?","holding":"The Ninth Circuit AFFIRMED the preliminary injunction, holding hiQ raised serious questions that the CFAA does not apply where data is publicly available without authentication ('the concept of authorization presupposes access restrictions'). Reaffirmed on remand after the Supreme Court GVR in light of Van Buren. Still a serious-questions / PI standard, not a full merits adjudication of the CFAA on a complete record.","instrument":"hiQ Labs, Inc. v. LinkedIn Corp., 938 F.3d 985 (9th Cir. 2019), vacated and remanded 141 S. Ct. 2752 (2021), adhered to on remand 31 F.4th 1180 (9th Cir. 2022)","decision_date":"2022-04-18","authority":"U.S. Court of Appeals for the Ninth Circuit (Judge Berzon)","status":"good-law","binding_scope":"Binds federal district courts within the Ninth Circuit only. Persuasive elsewhere. It is a PI affirmance, not a merits ruling.","applies_to":["public-data","technical-measures"],"practical_rule":"In the Ninth Circuit, scraping public, non-authenticated pages is unlikely to be a CFAA violation — but this says nothing about contract, copyright or trespass.","source_type":"court-opinion","notes":"Widely mis-cited as 'scraping is legal.' It only addressed the CFAA, only at the PI stage."}},{"id":"hiq-labs-inc-v-linkedin-corp-no-17-cv-03301-emc-","dataset":"scraping","jurisdiction":"US federal (N.D. Cal.)","instrument":"hiQ Labs, Inc. v. LinkedIn Corp., No. 17-cv-03301-EMC (N.D. Cal. Nov. 4, 2022) (summary judgment); stipulated judgment Dec. 2022","status":"in-force","summary":"Can a scraper be liable for breach of a website's User Agreement even if it wins the CFAA question? — Yes. On remand, the N.D. Cal. granted LinkedIn summary judgment that hiQ BREACHED LinkedIn's User Agreement (which prohibits scraping and automated access), and rejected hiQ's unconscionability defense. hiQ prevailed on the CFAA at the PI stage but LOST on contract. The case then resolved by a December 2022 stipulated judgment / permanent injunction against hiQ.","source_url":"https://caselaw.findlaw.com/court/us-dis-crt-n-d-cal/2182242.html","source_status":"blocked","not_live_law":false,"detail":{"question":"Can a scraper be liable for breach of a website's User Agreement even if it wins the CFAA question?","holding":"Yes. On remand, the N.D. Cal. granted LinkedIn summary judgment that hiQ BREACHED LinkedIn's User Agreement (which prohibits scraping and automated access), and rejected hiQ's unconscionability defense. hiQ prevailed on the CFAA at the PI stage but LOST on contract. The case then resolved by a December 2022 stipulated judgment / permanent injunction against hiQ.","instrument":"hiQ Labs, Inc. v. LinkedIn Corp., No. 17-cv-03301-EMC (N.D. Cal. Nov. 4, 2022) (summary judgment); stipulated judgment Dec. 2022","decision_date":"2022-11-04","authority":"U.S. District Court, N.D. Cal. (Judge Edward Chen)","status":"good-law","binding_scope":"District-court ruling; persuasive only. Illustrates that contract liability is independent of the CFAA.","applies_to":["contract-terms","public-data","logged-in-data"],"practical_rule":"Winning on the CFAA does not defeat a breach-of-contract claim; a User Agreement banning scraping can still be enforced against you.","source_type":"court-opinion","notes":"The single most important corrective to the 'hiQ made scraping legal' myth: hiQ ultimately lost."}},{"id":"van-buren-v-united-states-593-u-s-374-141-s-ct-1","dataset":"scraping","jurisdiction":"US (nationwide)","instrument":"Van Buren v. United States, 593 U.S. 374, 141 S. Ct. 1648 (2021); 18 U.S.C. § 1030(a)(2), (e)(6)","status":"in-force","summary":"When does someone 'exceed authorized access' under the CFAA? — MERITS HOLDING (6-3): An individual 'exceeds authorized access' only when he accesses areas of a computer (files, folders, databases) that are off-limits to him — a 'gates-up-or-down' inquiry. Using authorized access for an improper purpose or in violation of use policies is NOT a CFAA violation. The Court expressly reserved (footnote 8) whether 'authorization' turns only on technological limits or also on contract/policy limits.","source_url":"https://www.supremecourt.gov/opinions/20pdf/19-783_k53l.pdf","source_status":"verified","not_live_law":false,"detail":{"question":"When does someone 'exceed authorized access' under the CFAA?","holding":"MERITS HOLDING (6-3): An individual 'exceeds authorized access' only when he accesses areas of a computer (files, folders, databases) that are off-limits to him — a 'gates-up-or-down' inquiry. Using authorized access for an improper purpose or in violation of use policies is NOT a CFAA violation. The Court expressly reserved (footnote 8) whether 'authorization' turns only on technological limits or also on contract/policy limits.","instrument":"Van Buren v. United States, 593 U.S. 374, 141 S. Ct. 1648 (2021); 18 U.S.C. § 1030(a)(2), (e)(6)","decision_date":"2021-06-03","authority":"Supreme Court of the United States (Justice Barrett)","status":"good-law","binding_scope":"Binding on all US federal and state courts. Resolved the circuit split with the First, Fifth, Seventh and Eleventh Circuits' older broad readings.","applies_to":["technical-measures","public-data","logged-in-data","contract-terms"],"practical_rule":"Breaching terms of use is not, by itself, a federal computer crime; CFAA turns on whether you entered an area you were not permitted to enter.","source_type":"court-opinion","notes":"Did NOT legalize ToS violations — it left the contract-vs-code question open and does not touch contract, copyright, or trespass claims."}},{"id":"facebook-inc-v-power-ventures-inc-844-f-3d-1058-","dataset":"scraping","jurisdiction":"US federal (Ninth Circuit)","instrument":"Facebook, Inc. v. Power Ventures, Inc., 844 F.3d 1058 (9th Cir. 2016)","status":"in-force","summary":"Does a cease-and-desist letter revoking access create CFAA liability for continued scraping? — MERITS HOLDING: Power initially had authorization (via users), but once Facebook sent a cease-and-desist letter and blocked its IP, Power's continued access was 'without authorization' and violated the CFAA and Cal. Penal Code § 502. The court stressed that a bare terms-of-use violation, 'without more,' is NOT enough — the operative event was explicit revocation. Bypassing an IP block alone is not enough either.","source_url":"https://cdn.ca9.uscourts.gov/datastore/opinions/2016/07/12/13-17102.pdf","source_status":"verified","not_live_law":false,"detail":{"question":"Does a cease-and-desist letter revoking access create CFAA liability for continued scraping?","holding":"MERITS HOLDING: Power initially had authorization (via users), but once Facebook sent a cease-and-desist letter and blocked its IP, Power's continued access was 'without authorization' and violated the CFAA and Cal. Penal Code § 502. The court stressed that a bare terms-of-use violation, 'without more,' is NOT enough — the operative event was explicit revocation. Bypassing an IP block alone is not enough either.","instrument":"Facebook, Inc. v. Power Ventures, Inc., 844 F.3d 1058 (9th Cir. 2016)","decision_date":"2016-12-09","authority":"U.S. Court of Appeals for the Ninth Circuit","status":"good-law","binding_scope":"Binds federal courts in the Ninth Circuit; persuasive elsewhere. Cert denied (2017). Consistent with Van Buren.","applies_to":["logged-in-data","public-data","technical-measures","contract-terms"],"practical_rule":"Once a site owner sends you a cease-and-desist and/or blocks you, continuing to access it can be a CFAA violation — a C&D letter has real legal force.","source_type":"court-opinion","notes":"Directly refutes 'a cease-and-desist letter has no legal effect.'"}},{"id":"united-states-v-nosal-nosal-i-676-f-3d-854-9th-c","dataset":"scraping","jurisdiction":"US federal (Ninth Circuit)","instrument":"United States v. Nosal (Nosal I), 676 F.3d 854 (9th Cir. 2012) (en banc); 18 U.S.C. § 1030","status":"in-force","summary":"Does violating an employer/website use policy 'exceed authorized access'? — MERITS HOLDING (en banc): 'Exceeds authorized access' is limited to ACCESS restrictions, not USE restrictions. Employees authorized to access data who then misuse it in violation of company policy do not violate the CFAA. Terms-of-use violations, without more, are not CFAA violations.","source_url":"https://www.courtlistener.com/opinion/625613/united-states-v-nosal/","source_status":"verified","not_live_law":false,"detail":{"question":"Does violating an employer/website use policy 'exceed authorized access'?","holding":"MERITS HOLDING (en banc): 'Exceeds authorized access' is limited to ACCESS restrictions, not USE restrictions. Employees authorized to access data who then misuse it in violation of company policy do not violate the CFAA. Terms-of-use violations, without more, are not CFAA violations.","instrument":"United States v. Nosal (Nosal I), 676 F.3d 854 (9th Cir. 2012) (en banc); 18 U.S.C. § 1030","decision_date":"2012-04-10","authority":"U.S. Court of Appeals for the Ninth Circuit (en banc)","status":"good-law","binding_scope":"Binds the Ninth Circuit; anticipated and is consistent with Van Buren.","applies_to":["contract-terms","logged-in-data"],"practical_rule":"A policy or ToS breach by an authorized user is not a federal crime in itself.","source_type":"court-opinion","notes":"Primary URL is a CourtListener locator; confirm the specific opinion page. Foundational to the narrow CFAA reading later adopted in Van Buren."}},{"id":"united-states-v-nosal-nosal-ii-844-f-3d-1024-9th","dataset":"scraping","jurisdiction":"US federal (Ninth Circuit)","instrument":"United States v. Nosal (Nosal II), 844 F.3d 1024 (9th Cir. 2016)","status":"in-force","summary":"Is using a current user's password after your own access was revoked a CFAA violation? — MERITS HOLDING: Where a company had affirmatively revoked the defendant's authorization, having former colleagues use a current employee's valid password to get in constituted access 'without authorization' under § 1030(a)(4). Distinguished from Nosal I (which concerned 'exceeds authorized access' by authorized users).","source_url":"https://www.courtlistener.com/opinion/4258547/united-states-v-nosal/","source_status":"verified","not_live_law":false,"detail":{"question":"Is using a current user's password after your own access was revoked a CFAA violation?","holding":"MERITS HOLDING: Where a company had affirmatively revoked the defendant's authorization, having former colleagues use a current employee's valid password to get in constituted access 'without authorization' under § 1030(a)(4). Distinguished from Nosal I (which concerned 'exceeds authorized access' by authorized users).","instrument":"United States v. Nosal (Nosal II), 844 F.3d 1024 (9th Cir. 2016)","decision_date":"2016-12-08","authority":"U.S. Court of Appeals for the Ninth Circuit","status":"good-law","binding_scope":"Binds the Ninth Circuit; cert denied. A vigorous dissent warned about criminalizing password sharing.","applies_to":["logged-in-data","technical-measures"],"practical_rule":"Do not use someone else's credentials to reach a system you have been banned from — that is access 'without authorization.'","source_type":"court-opinion","notes":"Primary URL is a CourtListener locator; confirm the opinion page."}},{"id":"ryanair-dac-v-booking-holdings-inc-no-1-20-cv-01","dataset":"scraping","jurisdiction":"US federal (D. Del.)","instrument":"Ryanair DAC v. Booking Holdings Inc., No. 1:20-cv-01191 (D. Del.); 18 U.S.C. § 1030","status":"in-force","summary":"Can scraping a password-protected portion of a site be a CFAA violation even after Van Buren — and does a jury verdict finding that survive post-trial review? — A Delaware jury (trial July 15-18, 2024) found Booking.com vicariously liable under the CFAA for accessing the password-protected 'myRyanair' area via a third party, with 'intent to defraud,' awarding exactly $5,000 (the statutory loss minimum). CRITICALLY, the trial court (Judge Bryson, Fed. Cir. by designation) LATER OVERTURNED the verdict on JMOL on January 22, 2025, finding Ryanair 'did not prove at least $5,000 in losses'; the matter is on appeal to the Third Circuit.","source_url":"https://corporate.ryanair.com/news/ryanair-wins-its-case-against-travel-giant-booking-com-in-delaware-court-ruling/","source_status":"verified","not_live_law":false,"detail":{"question":"Can scraping a password-protected portion of a site be a CFAA violation even after Van Buren — and does a jury verdict finding that survive post-trial review?","holding":"A Delaware jury (trial July 15-18, 2024) found Booking.com vicariously liable under the CFAA for accessing the password-protected 'myRyanair' area via a third party, with 'intent to defraud,' awarding exactly $5,000 (the statutory loss minimum). CRITICALLY, the trial court (Judge Bryson, Fed. Cir. by designation) LATER OVERTURNED the verdict on JMOL on January 22, 2025, finding Ryanair 'did not prove at least $5,000 in losses'; the matter is on appeal to the Third Circuit.","instrument":"Ryanair DAC v. Booking Holdings Inc., No. 1:20-cv-01191 (D. Del.); 18 U.S.C. § 1030","decision_date":"2025-01-22","authority":"U.S. District Court, D. Del. (Judge Curtis Bryson, Fed. Cir. sitting by designation)","status":"narrowed","binding_scope":"District-court proceedings; not binding precedent and now on appeal. Persuasive on how CFAA applies to password-protected areas post-Van Buren.","applies_to":["logged-in-data","technical-measures"],"practical_rule":"Scraping behind a login/technical barrier can trigger the CFAA — but the plaintiff must prove ≥$5,000 in cognizable loss, which is a real hurdle.","source_type":"court-opinion","notes":"The source_url is Ryanair's press release (NOT primary); the verdict form and JMOL order are the primary documents — confirm via PACER/CourtListener docket 1:20-cv-01191. This case cuts against 'scraping is never a CFAA violation,' but the JMOL reversal shows the loss element is decisive."}},{"id":"sandvig-v-barr-451-f-supp-3d-73-d-d-c-2020","dataset":"scraping","jurisdiction":"US federal (D.D.C.)","instrument":"Sandvig v. Barr, 451 F. Supp. 3d 73 (D.D.C. 2020)","status":"in-force","summary":"Do terms-of-service violations create CFAA liability (research context)? — MERITS HOLDING (as-applied): Violating a website's terms of service does not, standing alone, constitute a CFAA violation. Researchers who violate ToS to conduct audit testing of public websites do not thereby commit a crime under § 1030.","source_url":"https://www.courtlistener.com/opinion/4739724/sandvig-v-barr/","source_status":"verified","not_live_law":false,"detail":{"question":"Do terms-of-service violations create CFAA liability (research context)?","holding":"MERITS HOLDING (as-applied): Violating a website's terms of service does not, standing alone, constitute a CFAA violation. Researchers who violate ToS to conduct audit testing of public websites do not thereby commit a crime under § 1030.","instrument":"Sandvig v. Barr, 451 F. Supp. 3d 73 (D.D.C. 2020)","decision_date":"2020-03-27","authority":"U.S. District Court, D.D.C.","status":"good-law","binding_scope":"District-court ruling; persuasive only. Consistent with Van Buren (decided the following year).","applies_to":["contract-terms","public-data"],"practical_rule":"Breaking a website's ToS to access public data is not a federal crime, but it can still be a breach of contract.","source_type":"court-opinion","notes":"Primary URL is a CourtListener locator; confirm the opinion page."}},{"id":"meta-platforms-inc-v-bright-data-ltd-no-3-23-cv-","dataset":"scraping","jurisdiction":"US federal (N.D. Cal.)","instrument":"Meta Platforms, Inc. v. Bright Data Ltd., No. 3:23-cv-00077-EMC (N.D. Cal. Jan. 23, 2024)","status":"in-force","summary":"Do a platform's Terms of Service bind a scraper that only scrapes public data while logged OUT? — MERITS HOLDING (summary judgment): Meta's Facebook/Instagram Terms could not be construed to prohibit logged-off scraping of publicly available data, because those terms govern a 'user's' use of the products, and Bright Data was not a 'user' when logged off. A survival clause purporting to bind Bright Data after account termination was unenforceable. Meta LOST its breach-of-contract claim as to logged-out scraping.","source_url":"https://digitalcommons.law.scu.edu/cgi/viewcontent.cgi?article=3834&context=historical","source_status":"verified","not_live_law":false,"detail":{"question":"Do a platform's Terms of Service bind a scraper that only scrapes public data while logged OUT?","holding":"MERITS HOLDING (summary judgment): Meta's Facebook/Instagram Terms could not be construed to prohibit logged-off scraping of publicly available data, because those terms govern a 'user's' use of the products, and Bright Data was not a 'user' when logged off. A survival clause purporting to bind Bright Data after account termination was unenforceable. Meta LOST its breach-of-contract claim as to logged-out scraping.","instrument":"Meta Platforms, Inc. v. Bright Data Ltd., No. 3:23-cv-00077-EMC (N.D. Cal. Jan. 23, 2024)","decision_date":"2024-01-23","authority":"U.S. District Court, N.D. Cal. (Judge Edward Chen)","status":"good-law","binding_scope":"District-court ruling; persuasive only. Turns heavily on the specific wording of Meta's terms.","applies_to":["public-data","contract-terms"],"practical_rule":"Logged-out scraping of public data may escape a ToS aimed at 'users,' but this depends entirely on the exact contract wording — many terms DO bind visitors.","source_type":"court-opinion","notes":"Source is a law-school reproduction of the opinion; the primary is the N.D. Cal. docket (ECF 181). Meta retained a tortious-interference theory. Contract wording is decisive — not a general rule that logged-out scraping is contract-free."}},{"id":"x-corp-v-bright-data-ltd-no-3-23-cv-03698-wha-n-","dataset":"scraping","jurisdiction":"US federal (N.D. Cal.)","instrument":"X Corp. v. Bright Data Ltd., No. 3:23-cv-03698-WHA (N.D. Cal. May 9, 2024)","status":"in-force","summary":"Can a platform use breach-of-contract/state-law claims to stop scraping and resale of public user content? — MERITS HOLDING (motion to dismiss): X's breach-of-contract and state tort claims (tortious interference, misappropriation, unjust enrichment) based on scraping/selling public content were dismissed as PREEMPTED by the Copyright Act under conflict preemption — enforcing them would let X build a 'private copyright system' over content it does not own. Access-based claims failed because X did not plausibly allege intrusion.","source_url":"https://www.casemine.com/judgement/us/664045f747417f327c946394","source_status":"verified","not_live_law":false,"detail":{"question":"Can a platform use breach-of-contract/state-law claims to stop scraping and resale of public user content?","holding":"MERITS HOLDING (motion to dismiss): X's breach-of-contract and state tort claims (tortious interference, misappropriation, unjust enrichment) based on scraping/selling public content were dismissed as PREEMPTED by the Copyright Act under conflict preemption — enforcing them would let X build a 'private copyright system' over content it does not own. Access-based claims failed because X did not plausibly allege intrusion.","instrument":"X Corp. v. Bright Data Ltd., No. 3:23-cv-03698-WHA (N.D. Cal. May 9, 2024)","decision_date":"2024-05-09","authority":"U.S. District Court, N.D. Cal. (Judge William Alsup)","status":"good-law","binding_scope":"District-court ruling; persuasive only, and its broad conflict-preemption theory is contested. Fact-specific to X's ToS (users retain content ownership).","applies_to":["public-data","contract-terms","copyrighted-content"],"practical_rule":"A platform may struggle to use contract/tort law to monopolize public user content it does not itself own — but this is a contested district-court theory, not settled law.","source_type":"court-opinion","notes":"Source is a case aggregator; primary is the N.D. Cal. order (2024 WL 2113target). The preemption reasoning is in tension with Meta v. Bright Data and with the general rule that contract claims usually survive preemption due to the 'extra element' of promise."}},{"id":"register-com-inc-v-verio-inc-356-f-3d-393-2d-cir","dataset":"scraping","jurisdiction":"US federal (Second Circuit)","instrument":"Register.com, Inc. v. Verio, Inc., 356 F.3d 393 (2d Cir. 2004)","status":"in-force","summary":"Is a browsewrap/terms-of-use agreement enforceable against an automated agent that repeatedly uses the service with knowledge of the terms? — MERITS HOLDING: Yes. Even though Register.com's terms appeared after each WHOIS query, Verio was bound because it used the system repeatedly with actual knowledge of the terms — repeated use with knowledge constitutes assent. The court also accepted a trespass-to-chattels theory based on burden to system capacity, affirming the preliminary injunction.","source_url":"https://caselaw.findlaw.com/court/us-2nd-circuit/1420249.html","source_status":"blocked","not_live_law":false,"detail":{"question":"Is a browsewrap/terms-of-use agreement enforceable against an automated agent that repeatedly uses the service with knowledge of the terms?","holding":"MERITS HOLDING: Yes. Even though Register.com's terms appeared after each WHOIS query, Verio was bound because it used the system repeatedly with actual knowledge of the terms — repeated use with knowledge constitutes assent. The court also accepted a trespass-to-chattels theory based on burden to system capacity, affirming the preliminary injunction.","instrument":"Register.com, Inc. v. Verio, Inc., 356 F.3d 393 (2d Cir. 2004)","decision_date":"2004-01-23","authority":"U.S. Court of Appeals for the Second Circuit","status":"good-law","binding_scope":"Binds the Second Circuit; influential nationwide on browsewrap-by-conduct.","applies_to":["contract-terms","technical-measures","public-data"],"practical_rule":"A bot operator that keeps hitting a site knowing its terms can be contractually bound even without clicking 'I agree.'","source_type":"court-opinion","notes":"Refutes 'if you don't log in / don't click, no contract is formed' as an absolute rule."}},{"id":"nguyen-v-barnes-noble-inc-763-f-3d-1171-9th-cir-","dataset":"scraping","jurisdiction":"US federal (Ninth Circuit)","instrument":"Nguyen v. Barnes & Noble Inc., 763 F.3d 1171 (9th Cir. 2014)","status":"in-force","summary":"Is a browsewrap agreement enforceable where the user had no notice and did not manifest assent? — MERITS HOLDING: A browsewrap agreement is unenforceable where the website does not give reasonably conspicuous notice of the terms and the user takes no affirmative action manifesting assent. A hyperlink to terms in the corner of a page, without more, is insufficient.","source_url":"https://cdn.ca9.uscourts.gov/datastore/opinions/2014/08/18/12-56628.pdf","source_status":"verified","not_live_law":false,"detail":{"question":"Is a browsewrap agreement enforceable where the user had no notice and did not manifest assent?","holding":"MERITS HOLDING: A browsewrap agreement is unenforceable where the website does not give reasonably conspicuous notice of the terms and the user takes no affirmative action manifesting assent. A hyperlink to terms in the corner of a page, without more, is insufficient.","instrument":"Nguyen v. Barnes & Noble Inc., 763 F.3d 1171 (9th Cir. 2014)","decision_date":"2014-08-18","authority":"U.S. Court of Appeals for the Ninth Circuit","status":"good-law","binding_scope":"Binds the Ninth Circuit; widely followed on browsewrap enforceability.","applies_to":["contract-terms"],"practical_rule":"Whether ToS bind you depends on notice and assent; a buried browsewrap link may not be enforceable — but conspicuous terms plus repeated knowing use often are.","source_type":"court-opinion","notes":"Enforceability is fact-specific and turns on notice/assent, not on a bright line."}},{"id":"specht-v-netscape-communications-corp-306-f-3d-1","dataset":"scraping","jurisdiction":"US federal (Second Circuit)","instrument":"Specht v. Netscape Communications Corp., 306 F.3d 17 (2d Cir. 2002)","status":"in-force","summary":"Does clicking 'download' bind a user to terms located below the download button? — MERITS HOLDING: No. Where the terms were on a submerged portion of the screen not visible without scrolling, and users could download without seeing or assenting to them, no contract (and no arbitration clause) was formed. Assent requires reasonable notice.","source_url":"https://caselaw.findlaw.com/court/us-2nd-circuit/1055325.html","source_status":"blocked","not_live_law":false,"detail":{"question":"Does clicking 'download' bind a user to terms located below the download button?","holding":"MERITS HOLDING: No. Where the terms were on a submerged portion of the screen not visible without scrolling, and users could download without seeing or assenting to them, no contract (and no arbitration clause) was formed. Assent requires reasonable notice.","instrument":"Specht v. Netscape Communications Corp., 306 F.3d 17 (2d Cir. 2002)","decision_date":"2002-10-01","authority":"U.S. Court of Appeals for the Second Circuit (opinion by then-Judge Sotomayor)","status":"good-law","binding_scope":"Binds the Second Circuit; leading authority on inquiry notice for online terms.","applies_to":["contract-terms"],"practical_rule":"Terms a user never sees and never assents to are not binding — but conspicuous, clearly-presented terms are a different matter.","source_type":"court-opinion","notes":"Foundational to the clickwrap-vs-browsewrap distinction."}},{"id":"ebay-inc-v-bidder-s-edge-inc-100-f-supp-2d-1058-","dataset":"scraping","jurisdiction":"US federal (N.D. Cal.)","instrument":"eBay, Inc. v. Bidder's Edge, Inc., 100 F. Supp. 2d 1058 (N.D. Cal. 2000)","status":"in-force","summary":"Can automated crawling of a server constitute trespass to chattels? — PRELIMINARY-INJUNCTION FINDING: The court granted a preliminary injunction barring Bidder's Edge from using automated robots to query eBay's servers, on a trespass-to-chattels theory, reasoning that unauthorized use consuming server capacity, plus eBay's right to exclude, supported relief. This was a likelihood-of-success finding, not a final merits judgment (the case settled).","source_url":"https://law.justia.com/cases/federal/district-courts/FSupp2/100/1058/2478126/","source_status":"blocked","not_live_law":false,"detail":{"question":"Can automated crawling of a server constitute trespass to chattels?","holding":"PRELIMINARY-INJUNCTION FINDING: The court granted a preliminary injunction barring Bidder's Edge from using automated robots to query eBay's servers, on a trespass-to-chattels theory, reasoning that unauthorized use consuming server capacity, plus eBay's right to exclude, supported relief. This was a likelihood-of-success finding, not a final merits judgment (the case settled).","instrument":"eBay, Inc. v. Bidder's Edge, Inc., 100 F. Supp. 2d 1058 (N.D. Cal. 2000)","decision_date":"2000-05-24","authority":"U.S. District Court, N.D. Cal. (Judge Ronald Whyte)","status":"narrowed","binding_scope":"District-court PI order; its 'mere use' theory was later disapproved by the California Supreme Court in Hamidi (see next row). Persuasive at most.","applies_to":["public-data","technical-measures"],"practical_rule":"Heavy automated querying that burdens a server can support a trespass claim — but California law now requires actual impairment, so mere access is not enough.","source_type":"court-opinion","notes":"Read together with Hamidi: eBay's broad reading of trespass no longer holds in California."}},{"id":"intel-corp-v-hamidi-30-cal-4th-1342-2003","dataset":"scraping","jurisdiction":"California (state)","instrument":"Intel Corp. v. Hamidi, 30 Cal. 4th 1342 (2003)","status":"in-force","summary":"Does trespass to chattels apply to electronic access that causes no actual damage? — MERITS HOLDING (4-3): Under California law, trespass to chattels does NOT lie for electronic contact that causes no actual or threatened injury to the computer system or the possessor's interest in it. The court declined to adopt a rule of 'computer server inviolability' and disapproved readings of eBay that treat mere unauthorized 'use' as trespass.","source_url":"https://law.justia.com/cases/california/supreme-court/4th/30/1342.html","source_status":"blocked","not_live_law":false,"detail":{"question":"Does trespass to chattels apply to electronic access that causes no actual damage?","holding":"MERITS HOLDING (4-3): Under California law, trespass to chattels does NOT lie for electronic contact that causes no actual or threatened injury to the computer system or the possessor's interest in it. The court declined to adopt a rule of 'computer server inviolability' and disapproved readings of eBay that treat mere unauthorized 'use' as trespass.","instrument":"Intel Corp. v. Hamidi, 30 Cal. 4th 1342 (2003)","decision_date":"2003-06-30","authority":"Supreme Court of California","status":"good-law","binding_scope":"Binding in California; highly influential nationally on the actual-impairment requirement.","applies_to":["technical-measures","public-data"],"practical_rule":"A trespass-to-chattels claim against a scraper requires proof of actual impairment or measurable harm to the system, not just unauthorized access.","source_type":"court-opinion","notes":"The key limit on the eBay/Bidder's Edge trespass theory."}},{"id":"thomson-reuters-enterprise-centre-gmbh-v-ross-in","dataset":"scraping","jurisdiction":"US federal (D. Del.)","instrument":"Thomson Reuters Enterprise Centre GmbH v. Ross Intelligence Inc., 765 F. Supp. 3d 382 (D. Del. 2025); appeal 3d Cir. No. 25-2153","status":"in-force","summary":"Is copying copyrighted material to train a non-generative AI tool a fair use? — MERITS HOLDING (summary judgment): No. Judge Bibas held Ross infringed 2,243 original Westlaw headnotes and rejected fair use, concluding Ross 'took the headnotes to make it easier to develop a competing legal research tool.' The court emphasized this was NOT generative AI and that use was commercial, non-transformative, and market-substituting. Certified for interlocutory appeal.","source_url":"https://copyrightalliance.org/wp-content/uploads/2025/06/Bartz-v.-Anthropic-Order.pdf","source_status":"verified","not_live_law":false,"detail":{"question":"Is copying copyrighted material to train a non-generative AI tool a fair use?","holding":"MERITS HOLDING (summary judgment): No. Judge Bibas held Ross infringed 2,243 original Westlaw headnotes and rejected fair use, concluding Ross 'took the headnotes to make it easier to develop a competing legal research tool.' The court emphasized this was NOT generative AI and that use was commercial, non-transformative, and market-substituting. Certified for interlocutory appeal.","instrument":"Thomson Reuters Enterprise Centre GmbH v. Ross Intelligence Inc., 765 F. Supp. 3d 382 (D. Del. 2025); appeal 3d Cir. No. 25-2153","decision_date":"2025-02-11","authority":"U.S. District Court, D. Del. (Judge Stephanos Bibas, 3d Cir. by designation)","status":"good-law","binding_scope":"District-court ruling, now on interlocutory appeal to the Third Circuit under 28 U.S.C. § 1292(b). Not binding precedent; will become the first federal appellate word on AI-training fair use.","applies_to":["copyrighted-content"],"practical_rule":"Copying protected content to build a product that competes with the source is unlikely to be fair use, especially for non-generative tools.","source_type":"court-opinion","notes":"The source_url is the Bartz order which quotes and distinguishes the Ross opinion; the primary Ross order is D. Del. No. 1:20-cv-613, ECF 770 (confirm via CourtListener). Bibas revised his own 2023 opinion that had sent fair use to a jury."}},{"id":"bartz-v-anthropic-pbc-no-3-24-cv-05417-n-d-cal-j","dataset":"scraping","jurisdiction":"US federal (N.D. Cal.)","instrument":"Bartz v. Anthropic PBC, No. 3:24-cv-05417 (N.D. Cal. June 23, 2025)","status":"in-force","summary":"Is training an LLM on copyrighted books fair use, and does the source of the copies matter? — MERITS HOLDING (partial summary judgment): (1) Training LLMs on books was 'exceedingly transformative' fair use; (2) format-shifting lawfully PURCHASED print books to digital for an internal library was fair use; BUT (3) downloading and retaining 7+ million PIRATED books to build a permanent central library was NOT fair use — that issue was set for trial. The class then settled.","source_url":"https://copyrightalliance.org/wp-content/uploads/2025/06/Bartz-v.-Anthropic-Order.pdf","source_status":"verified","not_live_law":false,"detail":{"question":"Is training an LLM on copyrighted books fair use, and does the source of the copies matter?","holding":"MERITS HOLDING (partial summary judgment): (1) Training LLMs on books was 'exceedingly transformative' fair use; (2) format-shifting lawfully PURCHASED print books to digital for an internal library was fair use; BUT (3) downloading and retaining 7+ million PIRATED books to build a permanent central library was NOT fair use — that issue was set for trial. The class then settled.","instrument":"Bartz v. Anthropic PBC, No. 3:24-cv-05417 (N.D. Cal. June 23, 2025)","decision_date":"2025-06-23","authority":"U.S. District Court, N.D. Cal. (Judge William Alsup)","status":"good-law","binding_scope":"District-court ruling; persuasive only. Addressed only inputs; the authors did not allege infringing outputs.","applies_to":["copyrighted-content"],"practical_rule":"Training on LAWFULLY acquired works has a strong fair-use case; building a library from pirated copies does not — provenance of your training data is legally decisive.","source_type":"court-opinion","notes":"The related settlement (next row) resolved the piracy liability before trial."}},{"id":"bartz-v-anthropic-pbc-no-3-24-cv-05417-n-d-cal-s","dataset":"scraping","jurisdiction":"US federal (N.D. Cal.)","instrument":"Bartz v. Anthropic PBC, No. 3:24-cv-05417 (N.D. Cal.) — settlement (prelim. approval Sept. 25, 2025; final approval July 20, 2026)","status":"unknown","summary":"What did the Bartz v. Anthropic settlement establish? — Not a merits ruling but a landmark class settlement: a $1.5 billion fund, granted FINAL APPROVAL on July 20, 2026 by Judge Araceli Martínez-Olguín, who held it 'provides substantial benefits to the Class in light of the novel claims asserted.' Described as 'the largest copyright class action settlement in US history,' paying about $3,100 per author for more than 480,000 works (440,490 of 482,460 eligible works claimed), covering books downloaded from LibGen and the Pirate Library Mirror.","source_url":"https://cdn.arstechnica.net/wp-content/uploads/2026/07/Bartz-v-Anthropic-Order-Approving-Settlement-7-20-26.pdf","source_status":"verified","not_live_law":false,"detail":{"question":"What did the Bartz v. Anthropic settlement establish?","holding":"Not a merits ruling but a landmark class settlement: a $1.5 billion fund, granted FINAL APPROVAL on July 20, 2026 by Judge Araceli Martínez-Olguín, who held it 'provides substantial benefits to the Class in light of the novel claims asserted.' Described as 'the largest copyright class action settlement in US history,' paying about $3,100 per author for more than 480,000 works (440,490 of 482,460 eligible works claimed), covering books downloaded from LibGen and the Pirate Library Mirror.","instrument":"Bartz v. Anthropic PBC, No. 3:24-cv-05417 (N.D. Cal.) — settlement (prelim. approval Sept. 25, 2025; final approval July 20, 2026)","decision_date":"2026-07-20","authority":"U.S. District Court, N.D. Cal. (Judge Araceli Martínez-Olguín)","status":"settled-before-judgment","binding_scope":"Binds only the settling parties/class; no precedential holding on the piracy liability. Signals large exposure for training on pirated corpora.","applies_to":["copyrighted-content"],"practical_rule":"Training on pirated corpora carries settlement/damages exposure in the billions even where the training use itself may be transformative.","source_type":"court-opinion","notes":"Source is a media-hosted copy of the approval order; confirm via the N.D. Cal. docket. Releases input/acquisition claims through ~Aug. 25, 2025; does not release output claims."}},{"id":"kadrey-v-meta-platforms-inc-no-3-23-cv-03417-vc-","dataset":"scraping","jurisdiction":"US federal (N.D. Cal.)","instrument":"Kadrey v. Meta Platforms, Inc., No. 3:23-cv-03417-VC (N.D. Cal. June 25, 2025)","status":"in-force","summary":"Is training an LLM on copyrighted books fair use where plaintiffs argue market harm? — MERITS HOLDING (summary judgment for Meta) ON A NARROW GROUND: Judge Chhabria granted Meta summary judgment on fair use but stressed 'this ruling does not stand for the proposition that Meta's use of copyrighted materials to train its language models is lawful. It stands only for the proposition that these plaintiffs made the wrong arguments and failed to develop a record.' He called market effect the 'single most important' fair-use factor and warned (in dicta) that in many cases training that floods the market with competing works 'will likely be illegal.'","source_url":"https://www.courtlistener.com/docket/67569326/kadrey-v-meta-platforms-inc/","source_status":"verified","not_live_law":false,"detail":{"question":"Is training an LLM on copyrighted books fair use where plaintiffs argue market harm?","holding":"MERITS HOLDING (summary judgment for Meta) ON A NARROW GROUND: Judge Chhabria granted Meta summary judgment on fair use but stressed 'this ruling does not stand for the proposition that Meta's use of copyrighted materials to train its language models is lawful. It stands only for the proposition that these plaintiffs made the wrong arguments and failed to develop a record.' He called market effect the 'single most important' fair-use factor and warned (in dicta) that in many cases training that floods the market with competing works 'will likely be illegal.'","instrument":"Kadrey v. Meta Platforms, Inc., No. 3:23-cv-03417-VC (N.D. Cal. June 25, 2025)","decision_date":"2025-06-25","authority":"U.S. District Court, N.D. Cal. (Judge Vince Chhabria)","status":"good-law","binding_scope":"District-court ruling; persuasive only. Expressly limited to the record these 13 authors made.","applies_to":["copyrighted-content"],"practical_rule":"Do not read Kadrey as blessing AI training — a plaintiff who proves market dilution may well win, because the court flagged that as the decisive factor.","source_type":"court-opinion","notes":"Source is the CourtListener docket; confirm the SJ opinion (2025 WL 1752484). Reflects a doctrinal split with Alsup's transformative-use emphasis in Bartz."}},{"id":"getty-images-us-inc-v-stability-ai-ltd-2025-ewhc","dataset":"scraping","jurisdiction":"United Kingdom (England & Wales)","instrument":"Getty Images (US) Inc. v. Stability AI Ltd. [2025] EWHC 2863 (Ch)","status":"in-force","summary":"Does training a generative image model on scraped copyrighted images infringe UK copyright, and does distributing the model in the UK infringe? — MERITS HOLDING (Nov. 4, 2025): Getty ABANDONED its primary training/output copyright and database claims mid-trial (June 2025) for lack of evidence that training occurred in the UK. On the surviving claims: the secondary-infringement claim was REJECTED because the model weights 'are not themselves an infringing copy and they do not store an infringing copy'; the trademark findings were 'historic' and 'extremely limited,' with no infringement for Stable Diffusion XL, XL Turbo, or 1.6.","source_url":"https://www.judiciary.uk/judgments/getty-images-v-stability-ai/","source_status":"verified","not_live_law":false,"detail":{"question":"Does training a generative image model on scraped copyrighted images infringe UK copyright, and does distributing the model in the UK infringe?","holding":"MERITS HOLDING (Nov. 4, 2025): Getty ABANDONED its primary training/output copyright and database claims mid-trial (June 2025) for lack of evidence that training occurred in the UK. On the surviving claims: the secondary-infringement claim was REJECTED because the model weights 'are not themselves an infringing copy and they do not store an infringing copy'; the trademark findings were 'historic' and 'extremely limited,' with no infringement for Stable Diffusion XL, XL Turbo, or 1.6.","instrument":"Getty Images (US) Inc. v. Stability AI Ltd. [2025] EWHC 2863 (Ch)","decision_date":"2025-11-04","authority":"High Court of Justice (Chancery Division), Mrs Justice Joanna Smith DBE","status":"good-law","binding_scope":"Binding in England & Wales (High Court level; subject to appeal). Territorially specific: the court proceeded on the basis that training occurred outside the UK.","applies_to":["copyrighted-content","public-data","technical-measures"],"practical_rule":"A model trained abroad that does not store copies of the works is not, by mere distribution in the UK, a secondary-infringing 'article' — but reproducing trademarks/watermarks in outputs can still infringe.","source_type":"court-opinion","notes":"Confirm exact judiciary.uk URL/neutral citation [2025] EWHC 2863 (Ch). Gave little guidance on the core question of whether training on copyrighted works infringes, because of the jurisdiction/evidence issues."}},{"id":"the-new-york-times-co-v-microsoft-corp-openai-no","dataset":"scraping","jurisdiction":"US federal (S.D.N.Y.)","instrument":"The New York Times Co. v. Microsoft Corp. & OpenAI, No. 1:23-cv-11195 (S.D.N.Y. Apr. 4, 2025)","status":"in-force","summary":"Do copyright, contributory-infringement and DMCA claims against an AI developer survive a motion to dismiss? — MERITS HOLDING (motion to dismiss, April 4, 2025): Direct and contributory copyright infringement claims and trademark-dilution claims SURVIVED; the statute-of-limitations defense was rejected at the pleading stage. Common-law unfair-competition/misappropriation claims were dismissed with prejudice. Most DMCA § 1202(b) CMI claims were dismissed without prejudice (some survived against OpenAI in the Daily News and CIR actions). No fair-use ruling was made.","source_url":"https://law.justia.com/cases/federal/district-courts/new-york/nysdce/1:2023cv11195/612697/514/","source_status":"blocked","not_live_law":false,"detail":{"question":"Do copyright, contributory-infringement and DMCA claims against an AI developer survive a motion to dismiss?","holding":"MERITS HOLDING (motion to dismiss, April 4, 2025): Direct and contributory copyright infringement claims and trademark-dilution claims SURVIVED; the statute-of-limitations defense was rejected at the pleading stage. Common-law unfair-competition/misappropriation claims were dismissed with prejudice. Most DMCA § 1202(b) CMI claims were dismissed without prejudice (some survived against OpenAI in the Daily News and CIR actions). No fair-use ruling was made.","instrument":"The New York Times Co. v. Microsoft Corp. & OpenAI, No. 1:23-cv-11195 (S.D.N.Y. Apr. 4, 2025)","decision_date":"2025-04-04","authority":"U.S. District Court, S.D.N.Y. (Judge Sidney Stein)","status":"good-law","binding_scope":"District-court MTD ruling; not a merits determination of infringement or fair use. Persuasive only.","applies_to":["copyrighted-content","public-data"],"practical_rule":"Surviving a motion to dismiss is not a win on the merits — but the core infringement and contributory claims against AI developers are proceeding to discovery and trial.","source_type":"court-opinion","notes":"Consolidated with related author and Daily News/CIR actions. Diverges from some N.D. Cal. rulings by letting output-based user-infringement theories proceed."}},{"id":"raw-story-media-inc-v-openai-inc-no-1-24-cv-0151","dataset":"scraping","jurisdiction":"US federal (S.D.N.Y.)","instrument":"Raw Story Media, Inc. v. OpenAI, Inc., No. 1:24-cv-01514 (S.D.N.Y. Nov. 7, 2024); 17 U.S.C. § 1202(b)","status":"in-force","summary":"Can a publisher sue an AI company under DMCA § 1202(b) for stripping copyright-management information from training data? — MERITS HOLDING (motion to dismiss): The § 1202(b) claim was dismissed for lack of Article III standing. Removal of CMI before training, without any allegation that OpenAI DISSEMINATED the plaintiffs' articles stripped of CMI, is not a concrete injury-in-fact ('No concrete harm, no standing'). Dismissed without prejudice.","source_url":"https://patentlyo.com/media/2024/11/Raw-Story-Order.pdf","source_status":"verified","not_live_law":false,"detail":{"question":"Can a publisher sue an AI company under DMCA § 1202(b) for stripping copyright-management information from training data?","holding":"MERITS HOLDING (motion to dismiss): The § 1202(b) claim was dismissed for lack of Article III standing. Removal of CMI before training, without any allegation that OpenAI DISSEMINATED the plaintiffs' articles stripped of CMI, is not a concrete injury-in-fact ('No concrete harm, no standing'). Dismissed without prejudice.","instrument":"Raw Story Media, Inc. v. OpenAI, Inc., No. 1:24-cv-01514 (S.D.N.Y. Nov. 7, 2024); 17 U.S.C. § 1202(b)","decision_date":"2024-11-07","authority":"U.S. District Court, S.D.N.Y. (Judge Colleen McMahon)","status":"good-law","binding_scope":"District-court ruling; persuasive only. In tension with Intercept Media v. OpenAI (Judge Rakoff, Feb. 2025), which let a § 1202(b)(1) claim proceed.","applies_to":["copyrighted-content"],"practical_rule":"A bare CMI-removal claim without proof of dissemination or concrete harm is likely to fail on standing — but this is contested among SDNY judges.","source_type":"court-opinion","notes":"Source is a media-hosted copy of the order; confirm via CourtListener/PACER. Illustrates the split with N.D. Cal.'s substantive 'identicality' rationale (see next row)."}},{"id":"doe-1-v-github-inc-no-4-22-cv-06823-n-d-cal-17-u","dataset":"scraping","jurisdiction":"US federal (N.D. Cal.)","instrument":"Doe 1 v. GitHub, Inc., No. 4:22-cv-06823 (N.D. Cal.); 17 U.S.C. § 1202(b) (Copilot litigation)","status":"in-force","summary":"Does DMCA § 1202(b) require that CMI be removed from an identical copy? — MERITS HOLDING (motion to dismiss): The N.D. Cal. dismissed § 1202(b) CMI claims on the substantive ground that the statute requires removal of CMI from IDENTICAL copies of the work on which the copyright holder placed the CMI; substantially-similar copies do not satisfy the requirement. The plaintiffs took an interlocutory appeal to the Ninth Circuit on the identicality question.","source_url":"https://www.courtlistener.com/docket/62924975/doe-1-v-github-inc/","source_status":"verified","not_live_law":false,"detail":{"question":"Does DMCA § 1202(b) require that CMI be removed from an identical copy?","holding":"MERITS HOLDING (motion to dismiss): The N.D. Cal. dismissed § 1202(b) CMI claims on the substantive ground that the statute requires removal of CMI from IDENTICAL copies of the work on which the copyright holder placed the CMI; substantially-similar copies do not satisfy the requirement. The plaintiffs took an interlocutory appeal to the Ninth Circuit on the identicality question.","instrument":"Doe 1 v. GitHub, Inc., No. 4:22-cv-06823 (N.D. Cal.); 17 U.S.C. § 1202(b) (Copilot litigation)","decision_date":"2024-01-03","authority":"U.S. District Court, N.D. Cal. (Judge Jon Tigar)","status":"good-law","binding_scope":"District-court ruling; persuasive only, and on interlocutory appeal to the Ninth Circuit. Reflects a district split with the SDNY standing approach.","applies_to":["copyrighted-content"],"practical_rule":"A § 1202 CMI claim is vulnerable unless the CMI was stripped from an exact copy of your work.","source_type":"court-opinion","notes":"Source is the CourtListener docket; confirm the specific dismissal order. Date approximate to the relevant dismissal — verify exact order date."}},{"id":"field-v-google-inc-412-f-supp-2d-1106-d-nev-2006","dataset":"scraping","jurisdiction":"US federal (D. Nev.)","instrument":"Field v. Google Inc., 412 F. Supp. 2d 1106 (D. Nev. 2006)","status":"in-force","summary":"Does ignoring robots.txt or the absence of a robots.txt/meta-tag opt-out have legal consequences? — MERITS HOLDING (summary judgment): Google's caching of a site whose owner did NOT deploy a 'no-archive' meta-tag / robots exclusion was protected by an IMPLIED LICENCE and by ESTOPPEL, because the owner knew Google would treat the absence of the tag as permission and chose not to opt out. Also found fair use and DMCA § 512(b) cache safe harbor. Robots/meta tags mattered only as evidence of the owner's intent/consent — not as a freestanding legal rule.","source_url":"https://www.courtlistener.com/opinion/2408379/field-v-google-inc/","source_status":"verified","not_live_law":false,"detail":{"question":"Does ignoring robots.txt or the absence of a robots.txt/meta-tag opt-out have legal consequences?","holding":"MERITS HOLDING (summary judgment): Google's caching of a site whose owner did NOT deploy a 'no-archive' meta-tag / robots exclusion was protected by an IMPLIED LICENCE and by ESTOPPEL, because the owner knew Google would treat the absence of the tag as permission and chose not to opt out. Also found fair use and DMCA § 512(b) cache safe harbor. Robots/meta tags mattered only as evidence of the owner's intent/consent — not as a freestanding legal rule.","instrument":"Field v. Google Inc., 412 F. Supp. 2d 1106 (D. Nev. 2006)","decision_date":"2006-01-19","authority":"U.S. District Court, D. Nev. (Judge Robert Jones)","status":"good-law","binding_scope":"District-court ruling; persuasive only. The leading authority that robots/meta signals bear on implied licence and estoppel.","applies_to":["technical-measures","copyrighted-content","public-data"],"practical_rule":"robots.txt is not itself a law, but honoring or ignoring it is evidence of notice, intent and (non-)consent that can decide implied-licence and estoppel questions.","source_type":"court-opinion","notes":"No court in any jurisdiction has held robots.txt legally binding of its own force. This case is the closest: it treats the robots/meta signal as evidence, not as an independent legal command."}},{"id":"directive-eu-2019-790-dsm-directive-arts-3-and-4","dataset":"scraping","jurisdiction":"European Union (all Member States)","instrument":"Directive (EU) 2019/790 (DSM Directive) Arts. 3 and 4","status":"in-force","summary":"Does EU law permit text and data mining of copyrighted works for AI training, and can rightholders opt out? — STATUTORY INSTRUMENT: Article 3 creates a mandatory, non-overridable TDM exception for scientific research by research organisations and cultural-heritage institutions with lawful access. Article 4 creates a general TDM exception (including commercial) that applies only where (a) the user has LAWFUL ACCESS and (b) the rightholder has NOT expressly reserved rights (opted out) in an appropriate manner — for content made public online, in a machine-readable form (Art. 4(3)).","source_url":"https://eur-lex.europa.eu/eli/dir/2019/790/oj","source_status":"verified","not_live_law":false,"detail":{"question":"Does EU law permit text and data mining of copyrighted works for AI training, and can rightholders opt out?","holding":"STATUTORY INSTRUMENT: Article 3 creates a mandatory, non-overridable TDM exception for scientific research by research organisations and cultural-heritage institutions with lawful access. Article 4 creates a general TDM exception (including commercial) that applies only where (a) the user has LAWFUL ACCESS and (b) the rightholder has NOT expressly reserved rights (opted out) in an appropriate manner — for content made public online, in a machine-readable form (Art. 4(3)).","instrument":"Directive (EU) 2019/790 (DSM Directive) Arts. 3 and 4","decision_date":"2019-04-17","authority":"European Parliament and Council (transposed by Member States)","status":"good-law","binding_scope":"Binds all EU Member States via national transposition. Interpretation is subject to CJEU rulings (a Hungarian reference on AI training is pending, judgment not expected until ~2027).","applies_to":["copyrighted-content","public-data","contract-terms"],"practical_rule":"In the EU you may mine copyrighted works for training only if you have lawful access AND the rightholder has not opted out; the research exception (Art. 3) cannot be contracted away.","source_type":"legislative-text","notes":"The commercial exception (Art. 4) is NOT unconditional — the opt-out and lawful-access preconditions are central."}},{"id":"kneschke-v-laion-e-v-lg-hamburg-27-sept-2024-az-","dataset":"scraping","jurisdiction":"Germany","instrument":"Kneschke v. LAION e.V., LG Hamburg, 27 Sept. 2024, Az. 310 O 227/23; OLG Hamburg, 10 Dec. 2025, 5 U 104/24","status":"in-force","summary":"Do the EU TDM exceptions cover scraping and copying to build an AI training dataset, and how must an opt-out be expressed? — MERITS HOLDING: The LG Hamburg (Sept. 27, 2024) held LAION's reproduction of a photo to build the LAION-5B dataset was covered by the § 60d UrhG / Art. 3 scientific-research TDM exception (LAION being a non-profit whose free dataset serves research), and dismissed the claim. In OBITER it indicated an opt-out in natural-language website terms might suffice under Art. 4(3). On appeal, the OLG Hamburg (Dec. 10, 2025, 5 U 104/24) confirmed the result and held § 44b (Art. 4 general TDM) ALSO applied, addressing the machine-readability of opt-outs.","source_url":"https://www.euipo.europa.eu/en/law/recent-case-law/germany-hamburg-district-court-310-o-22723-laion-v-robert-kneschke","source_status":"verified","not_live_law":false,"detail":{"question":"Do the EU TDM exceptions cover scraping and copying to build an AI training dataset, and how must an opt-out be expressed?","holding":"MERITS HOLDING: The LG Hamburg (Sept. 27, 2024) held LAION's reproduction of a photo to build the LAION-5B dataset was covered by the § 60d UrhG / Art. 3 scientific-research TDM exception (LAION being a non-profit whose free dataset serves research), and dismissed the claim. In OBITER it indicated an opt-out in natural-language website terms might suffice under Art. 4(3). On appeal, the OLG Hamburg (Dec. 10, 2025, 5 U 104/24) confirmed the result and held § 44b (Art. 4 general TDM) ALSO applied, addressing the machine-readability of opt-outs.","instrument":"Kneschke v. LAION e.V., LG Hamburg, 27 Sept. 2024, Az. 310 O 227/23; OLG Hamburg, 10 Dec. 2025, 5 U 104/24","decision_date":"2025-12-10","authority":"Landgericht Hamburg; Oberlandesgericht Hamburg","status":"good-law","binding_scope":"German court decisions; the OLG appeal ruling carries more weight but is not CJEU-level. First European rulings on TDM exceptions for AI datasets; may influence but not bind other Member States.","applies_to":["copyrighted-content","public-data","contract-terms","technical-measures"],"practical_rule":"Creating an AI training dataset can fall within the EU TDM exceptions, but a valid rights reservation (ideally machine-readable) can defeat the commercial Art. 4 route.","source_type":"court-opinion","notes":"Source is EUIPO's case-law database (regulator/official). The contested reproductions were dataset-building, not the model-training step itself. A German-language primary judgment text should be cited where available."}},{"id":"edpb-opinion-28-2024-on-personal-data-processing","dataset":"scraping","jurisdiction":"European Union / EEA","instrument":"EDPB Opinion 28/2024 on personal data processing in the context of AI models (Art. 64(2) GDPR)","status":"in-force","summary":"Can personal data be scraped and used to train AI models under the GDPR, and are trained models 'anonymous'? — REGULATOR GUIDANCE (Opinion 28/2024, 17 Dec. 2024): (1) An AI model trained on personal data is NOT automatically anonymous — anonymity must be assessed case-by-case, requiring that both direct/probabilistic extraction and extraction via queries be insignificant. (2) Legitimate interest (Art. 6(1)(f)) CAN be a lawful basis for development/deployment, but only after a three-step legitimate-interest assessment. (3) Unlawful processing in the development phase can taint later deployment.","source_url":"https://www.edpb.europa.eu/our-work-tools/our-documents/opinion-board-art-64/opinion-282024-certain-data-protection-aspects_en","source_status":"verified","not_live_law":false,"detail":{"question":"Can personal data be scraped and used to train AI models under the GDPR, and are trained models 'anonymous'?","holding":"REGULATOR GUIDANCE (Opinion 28/2024, 17 Dec. 2024): (1) An AI model trained on personal data is NOT automatically anonymous — anonymity must be assessed case-by-case, requiring that both direct/probabilistic extraction and extraction via queries be insignificant. (2) Legitimate interest (Art. 6(1)(f)) CAN be a lawful basis for development/deployment, but only after a three-step legitimate-interest assessment. (3) Unlawful processing in the development phase can taint later deployment.","instrument":"EDPB Opinion 28/2024 on personal data processing in the context of AI models (Art. 64(2) GDPR)","decision_date":"2024-12-17","authority":"European Data Protection Board","status":"good-law","binding_scope":"Not legally binding of itself, but authoritative and steers all EEA supervisory authorities toward consistent GDPR application. Uses 'case by case' and 'may/might' heavily, leaving room for DPA divergence.","applies_to":["personal-data","public-data","copyrighted-content"],"practical_rule":"Scraping personal data for AI training needs a documented legitimate-interest assessment; you cannot assume a trained model is anonymous or that public data is fair game.","source_type":"regulator-guidance","notes":"Confirm exact EDPB URL. Requested by the Irish DPC; follows the ChatGPT Taskforce report (23 May 2024)."}},{"id":"information-commissioner-v-clearview-ai-inc-uk-u","dataset":"scraping","jurisdiction":"United Kingdom","instrument":"Information Commissioner v. Clearview AI Inc. (UK Upper Tribunal, 7 Oct. 2025); UK GDPR Arts. 2-3","status":"in-force","summary":"Does UK/EU data-protection law reach a foreign company that scrapes residents' data, even for foreign clients? — MERITS HOLDING (Upper Tribunal, 7 Oct. 2025): The ICO DID have jurisdiction to fine Clearview and issue an enforcement notice. The UT overturned the First-tier Tribunal, holding Clearview's processing 'related to the monitoring of behaviour' of UK residents and fell within the material and territorial scope of the UK GDPR (Art. 3(2)(b)), even though Clearview has no UK establishment and served only foreign law-enforcement clients. Remitted to the FTT for the substantive appeal; Clearview granted permission to appeal to the Court of Appeal (Dec. 19, 2025).","source_url":"https://ico.org.uk/about-the-ico/media-centre/news-and-blogs/2025/10/uk-upper-tribunal-hands-down-judgment-on-clearview-ai-inc/","source_status":"verified","not_live_law":false,"detail":{"question":"Does UK/EU data-protection law reach a foreign company that scrapes residents' data, even for foreign clients?","holding":"MERITS HOLDING (Upper Tribunal, 7 Oct. 2025): The ICO DID have jurisdiction to fine Clearview and issue an enforcement notice. The UT overturned the First-tier Tribunal, holding Clearview's processing 'related to the monitoring of behaviour' of UK residents and fell within the material and territorial scope of the UK GDPR (Art. 3(2)(b)), even though Clearview has no UK establishment and served only foreign law-enforcement clients. Remitted to the FTT for the substantive appeal; Clearview granted permission to appeal to the Court of Appeal (Dec. 19, 2025).","instrument":"Information Commissioner v. Clearview AI Inc. (UK Upper Tribunal, 7 Oct. 2025); UK GDPR Arts. 2-3","decision_date":"2025-10-07","authority":"Upper Tribunal (Administrative Appeals Chamber)","status":"good-law","binding_scope":"Binding in the UK on the jurisdiction question (subject to the granted Court of Appeal appeal). Realigns UK law with the EU GDPR's extraterritorial reach.","applies_to":["personal-data","public-data"],"practical_rule":"Scraping images/data of UK or EU residents subjects you to UK/EU GDPR regardless of where your servers or clients are — 'the servers are in the US' is not a defense.","source_type":"regulator-guidance","notes":"Source is the ICO's own statement (regulator); the primary is the Upper Tribunal judgment (confirm neutral citation). The Dutch DPA separately fined Clearview EUR 30.5m; Italian, French (CNIL) and Greek DPAs also issued penalties."}},{"id":"ryanair-ltd-v-pr-aviation-bv-case-c-30-14-cjeu-1","dataset":"scraping","jurisdiction":"European Union","instrument":"Ryanair Ltd v. PR Aviation BV, Case C-30/14 (CJEU, 15 Jan. 2015)","status":"in-force","summary":"Can a database owner with no IP right in the data still restrict scraping by contract? — MERITS HOLDING (CJEU preliminary ruling): The Database Directive's mandatory user rights (Arts. 6, 8) and the anti-contracting-out rule (Art. 15) apply ONLY to databases protected by copyright or the sui generis right. Where a database is NOT so protected, the Directive does not apply, so the owner is FREE to restrict use (including scraping) by contract, subject to national law.","source_url":"https://curia.europa.eu/juris/liste.jsf?num=C-30/14","source_status":"blocked","not_live_law":false,"detail":{"question":"Can a database owner with no IP right in the data still restrict scraping by contract?","holding":"MERITS HOLDING (CJEU preliminary ruling): The Database Directive's mandatory user rights (Arts. 6, 8) and the anti-contracting-out rule (Art. 15) apply ONLY to databases protected by copyright or the sui generis right. Where a database is NOT so protected, the Directive does not apply, so the owner is FREE to restrict use (including scraping) by contract, subject to national law.","instrument":"Ryanair Ltd v. PR Aviation BV, Case C-30/14 (CJEU, 15 Jan. 2015)","decision_date":"2015-01-15","authority":"Court of Justice of the European Union","status":"good-law","binding_scope":"Binding across the EU as an authoritative interpretation of the Database Directive.","applies_to":["contract-terms","public-data"],"practical_rule":"Even data with no copyright or database right can be locked down by enforceable website terms — contract is an independent basis for restricting scraping in the EU.","source_type":"court-opinion","notes":"The key EU authority that contract can restrict scraping where IP rights do not."}},{"id":"directive-96-9-ec-database-directive-arts-7-8-cj","dataset":"scraping","jurisdiction":"European Union","instrument":"Directive 96/9/EC (Database Directive) Arts. 7-8; CJEU C-203/02, C-202/12, C-762/19","status":"in-force","summary":"When does the sui generis database right prevent extraction/re-utilisation of scraped data? — STATUTORY INSTRUMENT (as interpreted by CJEU): Art. 7 grants the maker of a database a right to prevent extraction and/or re-utilisation of the whole or a substantial part where there was substantial investment in obtaining, verifying or presenting the contents; Art. 8 sets lawful-user rights. CJEU case law (British Horseracing Board C-203/02; Innoweb C-202/12; CV-Online v. Melons C-762/19) refines 'substantial investment' and when a dedicated meta-search engine 're-utilises' a database.","source_url":"https://eur-lex.europa.eu/eli/dir/1996/9/oj","source_status":"verified","not_live_law":false,"detail":{"question":"When does the sui generis database right prevent extraction/re-utilisation of scraped data?","holding":"STATUTORY INSTRUMENT (as interpreted by CJEU): Art. 7 grants the maker of a database a right to prevent extraction and/or re-utilisation of the whole or a substantial part where there was substantial investment in obtaining, verifying or presenting the contents; Art. 8 sets lawful-user rights. CJEU case law (British Horseracing Board C-203/02; Innoweb C-202/12; CV-Online v. Melons C-762/19) refines 'substantial investment' and when a dedicated meta-search engine 're-utilises' a database.","instrument":"Directive 96/9/EC (Database Directive) Arts. 7-8; CJEU C-203/02, C-202/12, C-762/19","decision_date":"1996-03-11","authority":"European Parliament and Council; CJEU","status":"good-law","binding_scope":"Binds all EU Member States. CV-Online v. Melons (2021) narrowed the right by requiring a harm-to-investment analysis balancing against free information flow.","applies_to":["public-data","copyrighted-content"],"practical_rule":"Scraping a 'substantial part' of a database that reflects substantial investment can infringe the sui generis right independently of copyright — check for it before bulk extraction.","source_type":"legislative-text","notes":"The investment must be in obtaining/verifying/presenting existing data, not in creating it (British Horseracing Board)."}},{"id":"regulation-eu-2024-1689-ai-act-art-53-1-c-d-appl","dataset":"scraping","jurisdiction":"European Union","instrument":"Regulation (EU) 2024/1689 (AI Act) Art. 53(1)(c)-(d); application dates Arts. 113(b), 111(3)","status":"in-force","summary":"What copyright obligations does the EU AI Act impose on general-purpose AI model providers, and from when? — STATUTORY INSTRUMENT: Art. 53(1)(c) requires GPAI model providers to put in place a policy to comply with EU copyright law, including to identify and comply with Art. 4(3) DSM rights reservations (opt-outs), 'including through state-of-the-art technologies.' Art. 53(1)(d) requires a publicly available, 'sufficiently detailed summary' of training content per an AI Office template. These GPAI obligations APPLY FROM 2 AUGUST 2025 (Art. 113(b)); pre-existing models have until 2 August 2027 to comply (Art. 111(3)).","source_url":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","source_status":"verified","not_live_law":false,"detail":{"question":"What copyright obligations does the EU AI Act impose on general-purpose AI model providers, and from when?","holding":"STATUTORY INSTRUMENT: Art. 53(1)(c) requires GPAI model providers to put in place a policy to comply with EU copyright law, including to identify and comply with Art. 4(3) DSM rights reservations (opt-outs), 'including through state-of-the-art technologies.' Art. 53(1)(d) requires a publicly available, 'sufficiently detailed summary' of training content per an AI Office template. These GPAI obligations APPLY FROM 2 AUGUST 2025 (Art. 113(b)); pre-existing models have until 2 August 2027 to comply (Art. 111(3)).","instrument":"Regulation (EU) 2024/1689 (AI Act) Art. 53(1)(c)-(d); application dates Arts. 113(b), 111(3)","decision_date":"2024-06-13","authority":"European Parliament and Council","status":"good-law","binding_scope":"Binds all EU Member States; the GPAI obligations are in force since 2 Aug. 2025. Training-content summary template published by the AI Office 24 July 2025; a voluntary GPAI Code of Practice was published 10 July 2025.","applies_to":["copyrighted-content","public-data","contract-terms"],"practical_rule":"If you offer a GPAI model in the EU, you must operate a copyright/opt-out compliance policy and publish a training-data summary — these duties are already live.","source_type":"legislative-text","notes":"CELEX 32024R1689, OJ L 2024/1689, 12.7.2024. The opt-out compliance duty gives Art. 4(3) DSM reservations cross-border effect for GPAI providers."}},{"id":"copyright-act-of-japan-art-30-4-aca-general-unde","dataset":"scraping","jurisdiction":"Japan","instrument":"Copyright Act of Japan, Art. 30-4; ACA 'General Understanding on AI and Copyright' (2024)","status":"in-force","summary":"Does Japanese copyright law permit using copyrighted works for AI training, including commercially? — STATUTORY INSTRUMENT (Art. 30-4, in force 2019): It is permitted to exploit a work, to the extent necessary, where the purpose is NOT to enjoy the expression — including for 'information analysis' such as AI training — and this applies to COMMERCIAL uses. Subject to a proviso: it does not apply where the use would 'unreasonably prejudice the interests of the copyright owner.' The Agency for Cultural Affairs' 2024 'General Understanding on AI and Copyright' interprets the enjoyment/proviso limits.","source_url":"https://www.bunka.go.jp/english/policy/copyright/pdf/94055801_01.pdf","source_status":"verified","not_live_law":false,"detail":{"question":"Does Japanese copyright law permit using copyrighted works for AI training, including commercially?","holding":"STATUTORY INSTRUMENT (Art. 30-4, in force 2019): It is permitted to exploit a work, to the extent necessary, where the purpose is NOT to enjoy the expression — including for 'information analysis' such as AI training — and this applies to COMMERCIAL uses. Subject to a proviso: it does not apply where the use would 'unreasonably prejudice the interests of the copyright owner.' The Agency for Cultural Affairs' 2024 'General Understanding on AI and Copyright' interprets the enjoyment/proviso limits.","instrument":"Copyright Act of Japan, Art. 30-4; ACA 'General Understanding on AI and Copyright' (2024)","decision_date":"2019-01-01","authority":"National Diet of Japan; Agency for Cultural Affairs (guidance)","status":"good-law","binding_scope":"Binding statute in Japan. The 2024 ACA document is guidance, not legally binding, but influential.","applies_to":["copyrighted-content","public-data"],"practical_rule":"Japan has one of the most permissive AI-training regimes, but the exception fails if the use 'enjoys' the expression or unreasonably harms the rightholder's market.","source_type":"regulator-guidance","notes":"Source is the ACA's official English overview. The statutory text itself should be cited from the Japanese e-Gov statute portal for the primary instrument."}},{"id":"shanghai-character-license-administrative-co-v-a","dataset":"scraping","jurisdiction":"China","instrument":"Shanghai Character License Administrative Co. v. [AI company], Guangzhou Internet Court, (2024) Yue 0192 Min Chu No. 113","status":"in-force","summary":"Can a generative-AI service provider be liable for infringing outputs derived from training on a protected character? — MERITS HOLDING (first-instance): The Guangzhou Internet Court found the AIGC platform liable for infringing the reproduction and adaptation rights in the Ultraman works, because its outputs partially or wholly reproduced the original artistic expression. It ordered the provider to stop generating infringing images and pay RMB 10,000, and found the provider failed a reasonable duty of care under China's 2023 Interim Measures for Generative AI.","source_url":"https://www.twobirds.com/en/insights/2024/china/liability-of-ai-service-providers-for-copyright-infringement","source_status":"verified","not_live_law":false,"detail":{"question":"Can a generative-AI service provider be liable for infringing outputs derived from training on a protected character?","holding":"MERITS HOLDING (first-instance): The Guangzhou Internet Court found the AIGC platform liable for infringing the reproduction and adaptation rights in the Ultraman works, because its outputs partially or wholly reproduced the original artistic expression. It ordered the provider to stop generating infringing images and pay RMB 10,000, and found the provider failed a reasonable duty of care under China's 2023 Interim Measures for Generative AI.","instrument":"Shanghai Character License Administrative Co. v. [AI company], Guangzhou Internet Court, (2024) Yue 0192 Min Chu No. 113","decision_date":"2024-02-08","authority":"Guangzhou Internet Court","status":"good-law","binding_scope":"First-instance Chinese Internet Court judgment; not binding precedent in the common-law sense but influential as China's first AIGC output-infringement ruling.","applies_to":["copyrighted-content"],"practical_rule":"Deploying a generative model that outputs recognizable copyrighted characters can create liability for the service provider, plus AI-specific regulatory duties in China.","source_type":"court-opinion","notes":"Source is a law-firm analysis reproducing the ruling; the primary is the Guangzhou Internet Court judgment (2024) Yue 0192 Min Chu 113 (Chinese-language). Concerns outputs, not the legality of training-data acquisition."}},{"id":"li-v-liu-beijing-internet-court-2023-jing-0491-m","dataset":"scraping","jurisdiction":"China","instrument":"Li v. Liu, Beijing Internet Court, (2023) Jing 0491 Min Chu No. 11279","status":"in-force","summary":"Is an AI-generated image copyrightable, and who owns it? — MERITS HOLDING (first-instance): An image generated with Stable Diffusion through the user's iterative prompt selection and parameter adjustments IS a copyrightable work, because it reflected the user's 'aesthetic choices and personal judgment' (sufficient originality and human intellectual investment). The human USER, not the AI, is the author/owner. The defendant who reposted it (removing the watermark) infringed attribution and network-dissemination rights; ordered to apologize and pay RMB 500.","source_url":"https://www.twobirds.com/en/insights/2024/china/liability-of-ai-service-providers-for-copyright-infringement","source_status":"verified","not_live_law":false,"detail":{"question":"Is an AI-generated image copyrightable, and who owns it?","holding":"MERITS HOLDING (first-instance): An image generated with Stable Diffusion through the user's iterative prompt selection and parameter adjustments IS a copyrightable work, because it reflected the user's 'aesthetic choices and personal judgment' (sufficient originality and human intellectual investment). The human USER, not the AI, is the author/owner. The defendant who reposted it (removing the watermark) infringed attribution and network-dissemination rights; ordered to apologize and pay RMB 500.","instrument":"Li v. Liu, Beijing Internet Court, (2023) Jing 0491 Min Chu No. 11279","decision_date":"2023-11-27","authority":"Beijing Internet Court","status":"good-law","binding_scope":"First-instance Chinese Internet Court judgment; influential but not binding precedent. Contrasts with the US approach (Thaler; US Copyright Office) denying protection to AI output lacking human authorship.","applies_to":["copyrighted-content"],"practical_rule":"In China, sufficiently human-directed AI output can be protected and owned by the user — a materially different position from the US.","source_type":"court-opinion","notes":"Source is secondary; the primary is the Beijing Internet Court judgment (Chinese-language). Included to show the divergent international position on AI-output copyrightability."}},{"id":"copyright-act-2021-singapore-ss-243-244","dataset":"scraping","jurisdiction":"Singapore","instrument":"Copyright Act 2021 (Singapore) ss. 243-244","status":"in-force","summary":"Does Singapore permit copying works for computational data analysis / AI training, and what is the access condition? — STATUTORY INSTRUMENT: Section 244 makes it a permitted use to copy a work for computational data analysis (defined in s. 243 to include using a work to improve a computer program, i.e., machine learning), for both commercial and non-commercial purposes, and it CANNOT be excluded by contract. But the user must have LAWFUL ACCESS to the copy from which the training copy is made — circumventing a paywall or breaching a database's terms to obtain access is not lawful access.","source_url":"https://sso.agc.gov.sg/Act/CA2021","source_status":"verified","not_live_law":false,"detail":{"question":"Does Singapore permit copying works for computational data analysis / AI training, and what is the access condition?","holding":"STATUTORY INSTRUMENT: Section 244 makes it a permitted use to copy a work for computational data analysis (defined in s. 243 to include using a work to improve a computer program, i.e., machine learning), for both commercial and non-commercial purposes, and it CANNOT be excluded by contract. But the user must have LAWFUL ACCESS to the copy from which the training copy is made — circumventing a paywall or breaching a database's terms to obtain access is not lawful access.","instrument":"Copyright Act 2021 (Singapore) ss. 243-244","decision_date":"2021-11-21","authority":"Parliament of Singapore","status":"good-law","binding_scope":"Binding statute in Singapore.","applies_to":["copyrighted-content","public-data","contract-terms","technical-measures"],"practical_rule":"You may train on works in Singapore, even commercially and despite contrary ToS — but only if your access to the source copy was lawful (no paywall circumvention).","source_type":"legislative-text","notes":"The lawful-access condition is the practical limit; unlike the EU there is no opt-out mechanism, and unlike the EU commercial exception it cannot be contracted out of."}},{"id":"century-21-canada-ltd-partnership-v-rogers-commu","dataset":"scraping","jurisdiction":"Canada (British Columbia)","instrument":"Century 21 Canada Ltd Partnership v. Rogers Communications Inc., 2011 BCSC 1196","status":"in-force","summary":"Is a browsewrap/terms-of-use agreement enforceable against a scraper in Canada, and does robots.txt matter? — MERITS HOLDING: The BC Supreme Court held Zoocasa/Rogers liable for BREACH OF CONTRACT for scraping Century 21's listings in violation of the website Terms of Use — the first Canadian decision to uphold a 'browse wrap' agreement, where sufficient notice plus continued use forms a binding contract without an 'I agree' click. Also found copyright infringement in copied descriptions and photos. Adherence/non-adherence to robots.txt was treated as relevant evidence in the fair-dealing analysis, not as an independent legal command.","source_url":"https://www.canlii.org/en/bc/bcsc/doc/2011/2011bcsc1196/2011bcsc1196.html","source_status":"blocked","not_live_law":false,"detail":{"question":"Is a browsewrap/terms-of-use agreement enforceable against a scraper in Canada, and does robots.txt matter?","holding":"MERITS HOLDING: The BC Supreme Court held Zoocasa/Rogers liable for BREACH OF CONTRACT for scraping Century 21's listings in violation of the website Terms of Use — the first Canadian decision to uphold a 'browse wrap' agreement, where sufficient notice plus continued use forms a binding contract without an 'I agree' click. Also found copyright infringement in copied descriptions and photos. Adherence/non-adherence to robots.txt was treated as relevant evidence in the fair-dealing analysis, not as an independent legal command.","instrument":"Century 21 Canada Ltd Partnership v. Rogers Communications Inc., 2011 BCSC 1196","decision_date":"2011-09-02","authority":"Supreme Court of British Columbia (Justice Punnett)","status":"good-law","binding_scope":"Binding in British Columbia; persuasive elsewhere in Canada. Leading Canadian authority on browsewrap enforceability.","applies_to":["contract-terms","copyrighted-content","technical-measures","public-data"],"practical_rule":"In Canada, conspicuous website terms can bind a scraper by conduct; robots.txt is evidence bearing on fair dealing, not a standalone prohibition.","source_type":"court-opinion","notes":"Reinforces that no jurisdiction treats robots.txt as legally binding of its own force; here it was one evidentiary factor."}},{"id":"copyright-designs-and-patents-act-1988-s-29a","dataset":"scraping","jurisdiction":"United Kingdom","instrument":"Copyright, Designs and Patents Act 1988, s. 29A","status":"in-force","summary":"Does UK copyright law permit text and data mining for AI training? — STATUTORY INSTRUMENT: Section 29A permits making copies for computational analysis (text and data analysis) ONLY for NON-COMMERCIAL research, and only with lawful access; contract terms preventing such copying are unenforceable. The proposed broader COMMERCIAL TDM exception was abandoned in 2023. The UK's 2024-2025 consultation on copyright and AI has not, as of mid-2026, produced an enacted broad commercial exception.","source_url":"https://www.legislation.gov.uk/ukpga/1988/48/section/29A","source_status":"verified","not_live_law":false,"detail":{"question":"Does UK copyright law permit text and data mining for AI training?","holding":"STATUTORY INSTRUMENT: Section 29A permits making copies for computational analysis (text and data analysis) ONLY for NON-COMMERCIAL research, and only with lawful access; contract terms preventing such copying are unenforceable. The proposed broader COMMERCIAL TDM exception was abandoned in 2023. The UK's 2024-2025 consultation on copyright and AI has not, as of mid-2026, produced an enacted broad commercial exception.","instrument":"Copyright, Designs and Patents Act 1988, s. 29A","decision_date":"2014-06-01","authority":"UK Parliament","status":"good-law","binding_scope":"Binding statute across the UK.","applies_to":["copyrighted-content","public-data","contract-terms"],"practical_rule":"In the UK there is NO general commercial TDM exception — training a commercial model on UK-copyrighted works is not covered by s. 29A.","source_type":"legislative-text","notes":"This narrow scope is a major contrast with Japan/Singapore and even the EU's Art. 4. The Data (Use and Access) Act 2025 addressed data-sharing/ADM but did not enact a broad commercial TDM exception; verify current status."}},{"id":"reddit-inc-v-anthropic-pbc-no-cgc-25-625892-cal-","dataset":"scraping","jurisdiction":"US (California state / S.D.N.Y.)","instrument":"Reddit, Inc. v. Anthropic PBC, No. CGC-25-625892 (Cal. Super.); Reddit, Inc. v. Perplexity AI, Inc. (S.D.N.Y., filed Oct. 22, 2025)","status":"unknown","summary":"Can rightholders use state-law and contract claims (not copyright) to stop AI companies scraping their platforms? — DEVELOPING LITIGATION (not yet merits-resolved): Reddit sued Anthropic (June 2025, California state court) and Perplexity plus scraping intermediaries (Oct. 2025, S.D.N.Y.) on breach-of-contract, unjust-enrichment, trespass-to-chattels and unfair-competition theories, DELIBERATELY pleading no copyright claim to avoid preemption. In the Anthropic case the N.D. Cal. reportedly held (Mar. 30, 2026) the state claims were NOT preempted and remanded to state court, endorsing the contract route.","source_url":"https://www.courtlistener.com/?q=Reddit+v+Anthropic","source_status":"verified","not_live_law":false,"detail":{"question":"Can rightholders use state-law and contract claims (not copyright) to stop AI companies scraping their platforms?","holding":"DEVELOPING LITIGATION (not yet merits-resolved): Reddit sued Anthropic (June 2025, California state court) and Perplexity plus scraping intermediaries (Oct. 2025, S.D.N.Y.) on breach-of-contract, unjust-enrichment, trespass-to-chattels and unfair-competition theories, DELIBERATELY pleading no copyright claim to avoid preemption. In the Anthropic case the N.D. Cal. reportedly held (Mar. 30, 2026) the state claims were NOT preempted and remanded to state court, endorsing the contract route.","instrument":"Reddit, Inc. v. Anthropic PBC, No. CGC-25-625892 (Cal. Super.); Reddit, Inc. v. Perplexity AI, Inc. (S.D.N.Y., filed Oct. 22, 2025)","decision_date":"2025-06-04","authority":"California Superior Court; U.S. District Court, S.D.N.Y.","status":"untested","binding_scope":"No merits rulings yet; the preemption/remand ruling is a procedural district-court decision. Watch as the leading test of the contract-not-copyright strategy.","applies_to":["contract-terms","public-data","technical-measures"],"practical_rule":"Platforms are increasingly suing scrapers on contract/state-law theories precisely because they survive copyright preemption — do not assume 'no copyright claim' means 'safe.'","source_type":"court-opinion","notes":"The Mar. 30, 2026 remand ruling and docket numbers should be confirmed via CourtListener/PACER; flagged as reported but not primary-verified in detail."}},{"id":"concord-music-group-v-anthropic-n-d-cal-umg-v-su","dataset":"scraping","jurisdiction":"US federal (various)","instrument":"Concord Music Group v. Anthropic (N.D. Cal.); UMG v. Suno (D. Mass.); UMG v. Udio (S.D.N.Y.); Disney/Universal v. Midjourney (C.D. Cal.)","status":"unknown","summary":"How are music- and image-industry AI training suits resolving (Suno, Udio, Midjourney, Concord)? — DEVELOPING/MIXED: In Concord Music Group v. Anthropic (N.D. Cal., Mar. 25, 2025) the court DENIED the publishers' preliminary injunction against training on lyrics (no irreparable harm; overbroad request) and dismissed contributory claims with leave to amend — a preliminary/pleading posture, not a merits fair-use ruling. UMG/RIAA suits against Suno (D. Mass.) and Udio (S.D.N.Y.) partially SETTLED with licensing deals in late 2025 (UMG-Udio ~Oct. 2025; Warner with both). Disney/Universal v. Midjourney (C.D. Cal., filed June 2025) is in early procedure with no merits ruling.","source_url":"https://www.courtlistener.com/?q=UMG+v+Suno","source_status":"verified","not_live_law":false,"detail":{"question":"How are music- and image-industry AI training suits resolving (Suno, Udio, Midjourney, Concord)?","holding":"DEVELOPING/MIXED: In Concord Music Group v. Anthropic (N.D. Cal., Mar. 25, 2025) the court DENIED the publishers' preliminary injunction against training on lyrics (no irreparable harm; overbroad request) and dismissed contributory claims with leave to amend — a preliminary/pleading posture, not a merits fair-use ruling. UMG/RIAA suits against Suno (D. Mass.) and Udio (S.D.N.Y.) partially SETTLED with licensing deals in late 2025 (UMG-Udio ~Oct. 2025; Warner with both). Disney/Universal v. Midjourney (C.D. Cal., filed June 2025) is in early procedure with no merits ruling.","instrument":"Concord Music Group v. Anthropic (N.D. Cal.); UMG v. Suno (D. Mass.); UMG v. Udio (S.D.N.Y.); Disney/Universal v. Midjourney (C.D. Cal.)","decision_date":"2025-03-25","authority":"Various U.S. District Courts","status":"untested","binding_scope":"No merits/appellate holdings on fair use in these matters; PI denials and settlements do not establish that training is lawful or unlawful.","applies_to":["copyrighted-content","public-data"],"practical_rule":"The music/image training questions remain unresolved on the merits — settlements reflect commercial risk allocation, not legal rules.","source_type":"court-opinion","notes":"Settlement particulars derive from trade press; confirm dockets. Included to map the unsettled frontier, not as merits authority."}}],"total":40,"coverage":{"corpus_total":40,"matched":40,"jurisdictions_in_corpus":["California (state)","Canada (British Columbia)","China","European Union","European Union (all Member States)","European Union / EEA","Germany","Japan","Singapore","US (California state / S.D.N.Y.)","US (nationwide)","US federal (D. Del.)","US federal (D. Nev.)","US federal (D.D.C.)","US federal (N.D. Cal.)","US federal (Ninth Circuit / N.D. Cal.)","US federal (Ninth Circuit)","US federal (S.D.N.Y.)","US federal (Second Circuit)","US federal (various)","United Kingdom","United Kingdom (England & Wales)"],"by_status":{"in-force":37,"unknown":3},"source_health":{"blocked":9,"verified":31},"dead_rules_included":0},"as_of":"2026-07-28","interpretation":"Reference data, not legal advice. Every row cites a primary source — legislative text, a regulator document, or a court opinion — and records whether that URL resolved when the corpus was built; \"blocked\" means the host refuses automated checkers, not that the citation is bad. Rows marked not_live_law were vacated, superseded, or never took effect, and are included deliberately: a rule wrongly believed to be in force is the most expensive error in this area. Verify against the cited source before relying on any row."},"_meta":{"request_id":"req_e80d2a4b21924c10b7c4a94c54e0385f","product":"legal/scraping","version":"1.0.0","receipt_id":null,"cost":{"quoted":0,"charged":0,"currency":"USD","display":"0.000000"},"cache":{"hit":true,"age_s":38724},"provenance":{"sourcing_class":"B","upstream":["internal:legal-corpus@2026-07"],"fetched_at":"2026-10-13T11:51:42.405Z","license":"proprietary"},"freshness":{"as_of":"2026-07-28","staleness_s":38724},"confidence":1}}