Issue 11
Viewpoint
lawsuit · United States
Anthropic asks court to rule lyrics training is fair use in dispute with music publishers
On October 21–22, Anthropic asked a US court to rule that AI training on song lyrics is fair use, while music publishers disputed its arguments on harm and purpose. According to Law360, the judge may regard the "mash up" use of the lyrics as fair use.
Who is affected and how
- lawyer
- The fair use arguments (harm, purpose of use) and the judge's remarks in the case should be built into strategy in similar disputes.
regulation · European Union
AI Act: what applies, what was delayed, and the Article 50 transparency obligations
The AI Act is in force, with only the timing of some obligations changed under the omnibus regulation; the high-risk rules (HR, credit scoring) have been postponed to December 2027. The AI literacy…
compliance specialistTransparency and AI literacy obligations must already be met; for high-risk systems, the new deadline is the basis for planning.corporate lawyerContracts should address both the changed deadline and the allocation of liability under Article 50.lawyerClients should also be advised to check that legal search tools and eDiscovery solutions comply.lawsuit · United StatesOwners of Elvis Presley catalogue sue Anthropic over 274 songs
On October 23, the owners of Elvis Presley's catalogue sued Anthropic for copyright infringement, alleging Presley songs were used to train Claude. Reports name 274 songs.
lawyerListing works song by song has become the standard structure of such complaints, which should be considered both when drafting claims and when preparing a defence.perCanadian news publishers sue Anthropic for copyright infringement in Ontario
Major Canadian news publishers have filed a copyright lawsuit against Anthropic in an Ontario court. The suit alleges the company used newspaper content to train its Claude chatbot; the news was…
lawyerThe Canadian (Ontario) copyright forum is a new arena for international AI litigation, and the suit is worth following.corporate lawyerIn content-licence and data-source contracts, the risk of training use should be weighed under Canadian jurisdiction as well.lawsuit · United StatesState attorneys general line up to sue OpenAI over ChatGPT safety and consumer protection
Eight state attorneys general filed a coordinated suit against OpenAI on 14 October, Iowa followed on 20 October, and Texas's attorney general sued both OpenAI and Sam Altman on 23 October over…
corporate lawyerAI providers' consumer disclosures and risk warnings need to be re-examined.compliance specialistChild protection and data protection controls must be defensible at state level too.lawyerState attorney general suits open a new category of AI product liability litigation.Hungary
lawRomania weighs restricting legal advice from AI platforms, with lawyers' backinglawyerA dispute over protecting the monopoly on legal advice could arise elsewhere too.citizenLegal advice from AI chatbots may become only partly available.compliance specialistProviders offering legal AI on the Romanian market should watch the bill's progress.regulationPolish AI regulator's decisions to be reviewed by SOKiK under new civil procedure sectionlawyerThe new procedural route for challenging Polish AI supervisory decisions in court must be understood.compliance specialistThe option of appealing authority decisions is part of compliance risk management.toolHungarian developer, 25, builds AI for legal and tax questionscitizenA domestic AI assistant for legal and tax questions has appeared.lawyerThe domestic arrival of legal AI aimed at laypeople also touches on the limits of advice.professionHungarian legal education adds AI regulation to its taskstoolWerdict, an AI-native toolkit for lawyers, launcheslawyerA new Hungarian legal AI tool is worth trying in daily workflows.toolJogkódex BH-chat uses AI to search more than 300.000 court decisionslawyerAI summaries and search results should be checked against the original decision.lawPoland begins work on law to criminalise erotic deepfakescitizenCriminal-law protection is being prepared for deepfake victims.lawyerA new criminal offence gives victims' representatives an additional legal avenue.
More key stories of the week
case · United States
AI hallucinations in court filings are slowing, but judges still sound alarms
Between 21 and 23 October, several US reports said the share of fabricated citations is declining slightly, yet courts' responses remain inconsistent. A Florida lawyer was referred for possible…
lawyerCitations in every AI-assisted filing must be verified, as disciplinary consequences are possible.corporate lawyerRisk from AI errors is also growing for local counsel.lawsuit · United StatesHundreds of authors file new copyright claims against Meta over AI training
On October 20, hundreds of authors filed new claims against Meta in a federal court in northern California, alleging their works were used without permission to train AI. Plaintiffs' law firms, which…
lawyerThe strategy of mass individual claims is an alternative to class actions, so it means a new client-acquisition and litigation model on the plaintiffs' side.ruling · United StatesAppeals court lifts block on Mississippi DEI ban and removes judge over AI errors
On October 22, the Fifth Circuit Court of Appeals lifted the block on Mississippi's DEI ban and reassigned the case to another judge, because the presiding judge had used AI in the ruling, with…
lawyerAI errors in a ruling can serve as grounds for challenging it on appeal.regulation · European UnionFirst harmonised standard supporting the AI Act published in the Official Journal
On 19 October, the first harmonised standard supporting compliance with the AI Act was published in the Official Journal of the European Union. The quality management standard is a tool for AI…
compliance specialistCompliance documentation can be built on the standard.corporate lawyerThe standard can serve as a basis for contractual compliance requirements.data protection · European UnionEU backs GDPR change letting AI use personal data without consent
The Council of the EU backed a GDPR amendment that would allow AI to use personal data without consent, and also extend the deadline for reporting data breaches. According to NOYB, lobby groups…
compliance specialistThe digital omnibus amendments mean legal bases and notification processes must be reassessed.corporate lawyerThe proposed amendments could change the legal basis for AI development that processes personal data.citizenThe role of consent may weaken when their data is used for AI training.law · United StatesCalifornia enacts laws regulating AI in the workplace and in legal practice
California has enacted broad AI laws covering employers' use of AI and limits on the use of AI by lawyers and arbitrators. The laws also affect eDiscovery teams.
lawyerNew obligations apply to AI use and must be built into everyday practice.compliance specialistNew California limits apply to employers' AI-based decisions.ruling · United StatesUS Supreme Court leaves in place ruling that bars AI copyright claims
According to reports on October 20, the Supreme Court left unchanged the decision that rules out copyright claims attributed to AI. Human authorship thus remains the starting point of US law.
lawyerClients must be told they cannot count on copyright protection for purely AI-generated output.corporate lawyerContracts should settle ownership of AI-produced content and the proof of human involvement.profession · European UnionSwedish legal AI startup Legora raises $550 million at $8.8 billion valuation
Swedish legal AI startup Legora raised $550 million in a Series E round at a valuation of $8.8 billion. The company is accelerating its US expansion; the news appeared between 19 and 21 October.
lawyerThe legal AI tools market is flush with capital, so law firms need to develop a vendor strategy.corporate lawyerIn-house legal departments should expect Legora's increasingly strong US presence.lawsuit · United StatesReservoir sues Suno and Anthropic over songs by Joni Mitchell and Madonna
On October 22, Reservoir filed suit against Suno and Anthropic over the unauthorised use of music, with songs by Joni Mitchell, Johnny Cash and Madonna among those named in the complaints. The…
lawsuit · United StatesUMG and Sony ask court to make DistroKid hand over AI music upload data in Suno case
On October 19, Universal Music Group and Sony Music asked a federal court to compel DistroKid to disclose data on uploads of AI-generated tracks. Separately, a dispute over a subpoena has emerged…
lawyerSeeking data from a third-party distributor may become a recurring evidence-gathering tool in similar cases.compliance specialistMusic distribution platforms must prepare in advance for courts being asked to order disclosure of their AI music records.law · United StatesUS states and Congress move on AI in political ads as spending grows
Minnesota has a law targeting AI deepfakes in political ads, and Rep. Shontel Brown has introduced a bill that would require labels on AI-generated campaign content. According to the Wesleyan Media…
lawyerIn campaign-law cases, a suit can be brought for violating deepfake rules, as the Moore campaign's challenge shows.compliance specialistLabelling obligations for AI content in campaigns differ from state to state.regulation · European UnionAI Act obligations for general-purpose AI models: biological risk and incident reporting
Under the AI Act, providers of general-purpose AI models with systemic risk must assess and mitigate biological risks. According to Lawfare, enforcement of the incident reporting obligation remains…
compliance specialistGPAI model providers must build risk management and incident reporting processes.lawyerFor providers of large models, the relevant voluntary code of practice can serve as a basis for compliance.In brief
lawAnthropic's leaked bill would let the US block AI models over 50-plus deaths or $1 billion in damagecompliance specialistFederal override of state incident-reporting rules could change compliance plans.corporate lawyerThe liability and threshold provisions may signal the direction of federal legislation.perCopyright suits against AI companies in the US reach 150lawyerThe market-effect theory, which is central to fair use outcomes, must be supported with evidence.corporate lawyerCase trackers (class action trackers) help assess risk in dealings with AI vendors.data protectionUK GDPR still applies to AI after the Data (Use and Access) Act, NeuralTrust sayscompliance specialistAutomated decision-making procedures need to be reviewed under the new rules.corporate lawyerWhere UK data is processed, the GDPR compliance of AI systems must still be demonstrable.perCharacter.AI bankruptcy and Disney deal open a new phase in chatbot lawsuitslawyerBankruptcy proceedings may become a new way of handling lawsuits and claims against AI companies.compliance specialistThe fate of user data, especially children's data, must be addressed in bankruptcy proceedings as well.citizenWhen a chatbot service shuts down, the protection of user data is in question.caseAI agent liability: who pays when an agent gets it wrong?corporate lawyerLiability should be allocated in contracts before agents are deployed.lawyerWithout settled case law, advising clients on risk takes on a larger role.professionAI takes over junior lawyers' routine work, raising questions about their traininglawyerFirms must plan how young lawyers will gain practical experience.regulationCanadian Judicial Council issues new guidance for judges on artificial intelligenceregulationSocialists push for EU-level rules on deepfakescompliance specialistPlatforms and AI providers must reckon with new EU protective measures.lawyerNew EU tools may become available for deepfake victims to enforce their rights.professionFreshfields partners explore spin-off of AI unit as email leakslawyerLarge firms' AI units could become independent businesses that attract investors.rulingMunich court: memorisation of song lyrics in language model is relevant under copyright lawlawyerThe German ruling may offer a rough benchmark for the copyright classification of memorisation in EU cases.corporate lawyerModel developers and companies deploying models must reckon with the risk posed by memorised content.lawDenmark plans to protect faces and voices from deepfakes through copyright lawlawyerAn EU member-state model could influence regulation in other countries.regulationLaw Commission outlines plan to overhaul product liability for AI, apps and algorithmscorporate lawyerThe liability risk of products containing AI needs to be reassessed.lawyerProduct liability claims may extend to digital products.toolHarvey adds integrations and Intake product for in-house legal teamslawyerNew fact-checking and research integrations are being added to the tools for litigation and matter handling.corporate lawyerIntake is used for first-line screening of contract requests and for review.toolHarvey, Legora, Microsoft and Thomson Reuters compete for the legal AI front doorlawyerLarge firms are adopting several platforms, so there is no single standard tool.corporate lawyerIn procurement, integration, source verifiability and a multi-platform strategy matter.professionCourts experiment with AI, but decisions stay with humanscomplianceMeta delays new business terms after publisher pushback over AI scrapingcorporate lawyerWhen platform contracts are amended, provisions on use for AI training need separate review.lawUK deepfake victim's campaign leads to change in the lawcitizenVictims of deepfake porn have gained stronger legal protection.lawyerThe amended rules affect how victims are represented.professionGenerative AI a systemic risk to legal sector, and insurers struggle to price itlawyerProfessional indemnity insurance terms need to be renegotiated to cover AI use.corporate lawyerThe growth in provider risk due to AI calls for contractual protection.regulationNew Jersey and Illinois move toward regulating AIcompliance specialistExpect a patchwork of state-by-state rules.corporate lawyerCompanies operating in several states must follow local AI rules.rulingThomson Reuters v. ROSS: Law Tech AI summarises Third Circuit rulinglawyerAn appellate-level decision in AI-training copyright cases can serve as a authoritative reference.perGoogle unlikely to get ex-NPR host's AI voice lawsuit dismissedlawyerIn voice-cloning cases, claims can survive early motions to dismiss.data protectionWhatsApp help pages told company's AI not to mention data protection authoritiescitizenThe AI assistant may not tell users about the option of complaining to a regulator.compliance specialistUser-facing information should mention data protection remedies as a matter of expectation.caseSpain election deepfake sparks legal dispute over AI political advertisinglawyerThe legal treatment of AI political advertising is still unsettled at member-state level.citizenProtecting the authenticity of election content is the subject of a legal dispute.caseBrazil's Enter case raises questions on governing AI-powered legal serviceslawyerThe regulation of AI-powered legal services is contested in other countries as well.compliance specialistFor cross-border legal AI services, local regulation is a risk.caseAnalyses examine AI training and copyright in India amid ANI v OpenAI disputelawyerIndia's copyright framework differs from the US fair use doctrine.regulationICMP says AI firms are trying to 'strong-arm' Australia on copyrightperOpenEvidence and Doximity settle AI legal disputelawyerThe settlement is an example of how AI disputes in healthcare can be resolved.data protectionGoogle changes privacy policy to allow AI search data to be used for targeted adscitizenAI search data may end up in advertising profiles, so it is worth checking settings.compliance specialistCompanies using Google AI services should consider new data-processing notices.regulationWhite House 'Super Intelligence' task force to hold first meetingcorporate lawyerA liability-centred federal approach also affects contractual risk management.compliance specialistThe federal oversight approach may affect both providers and users of frontier models.rulingIowa Supreme Court upholds video evidence despite defendant's deepfake claimlawyerThe defence must back a deepfake argument with evidence.regulationProposed EU KIDS Act targets online platforms, age verification and AI chatbotscompliance specialistAge-verification obligations for chatbots and platforms are expected.corporate lawyerDivided supervisory powers mean dealing with several authorities.