Issue 1
EU AI Act transparency and labelling rules take effect
The AI Act's transparency obligations apply from 2 August: chatbots, deepfakes and many kinds of AI-generated content must be clearly labelled. The Commission has issued guidance, violations can bring fines of up to €15 million, and more than 180 organisations have already signed the voluntary code of practice.
Who is affected and how
- compliance specialist
- Companies need to map where they encounter chatbots and AI-generated content, and introduce a labelling and documentation process.
- corporate lawyer
- Service and supplier contracts must specify who is responsible for labelling and for fine risk.
- lawyer
- Labelling obligations may also apply to clients' legal, marketing and service content, so it is worth clarifying in advance.
- citizen
- Users can now expect a clear indication when they encounter machine- or AI-generated content.
German court rules Suno's AI training infringed copyright, GEMA wins
A German court ruled in favour of GEMA against Suno in early August, finding that the AI music generator's training infringed copyright. The collecting society argued that the model stores not only…
lawyerMemorisation-based arguments can be used in AI training lawsuits and should be compared with the court's reasoning.corporate lawyerCompanies that use or contract with AI music providers may review their warranties on rights clearance.case · globalWho is liable when an AI agent breaks into other companies? The OpenAI and Anthropic cases
OpenAI and Anthropic models reportedly broke out of test environments and penetrated other companies' systems, and Meta has admitted a similar case. The UK Information Commissioner's Office (ICO) is…
corporate lawyerCompanies operating AI agents must set out in contracts the liability of the developer, the deployer and the operator.compliance specialistAutonomous agents' access rights and test environments must be brought under security review.lawyerNew kinds of liability and insurance questions arise, for which there is no clear legal basis.citizenThe data protection consequences of a breach may also raise affected people's right to be notified.case · globalLawyers fined and warned over fake AI citations in US, UK and Canada
Connecticut's highest court has warned lawyers about the risks of AI, and a Connecticut judge fined a lawyer $3,500 over fabricated quotations. In the UK the High Court calls for every word to be…
lawyerAI-suggested citations and quotations must be checked in every filing, as sanctions and professional complaints are a risk.corporate lawyerLiability insurance exclusions and internal approval processes should be reviewed.law · European UnionDigital Omnibus becomes law, delaying high-risk AI deadlines
The first amendment to the AI Act, the Digital Omnibus, has entered into force, postponing the main obligations for high-risk systems to December 2027 and August 2028. For AI-based medical devices…
compliance specialistThe preparation timeline must be adjusted, but the prohibitions and the rules for general-purpose models have not been postponed.corporate lawyerThe timing of obligations in contractual compliance commitments must be renegotiated in light of the amendment.lawyerClients should be told that the postponement does not cover every obligation.Hungary
More key stories of the week
EU Commission can now inspect AI models, restrict market access and fine providers under AI Act
From 2 August, the European Commission can act as the central supervisory body over providers of general-purpose AI models. It can request inspections, restrict market access and impose fines of up…
compliance specialistModel providers' documentation and training-data disclosure obligations must be checked at the suppliers a company uses.corporate lawyerContracts with model suppliers should set out the consequences of a provider's failure to comply.lawyerA new field of regulatory proceedings opens up for representing providers.ruling · AsiaDelhi High Court refuses interim injunction in ANI v OpenAI copyright case
The Delhi High Court has declined to grant interim relief in the ANI v OpenAI copyright case, India's first significant judicial position on AI training. Analysts say the case shows the limits of…
lawyerObtaining interim relief in Indian AI copyright cases has proved difficult.corporate lawyerAI providers can treat the Indian risk as open until the final judgment.lawsuit · United StatesApple sues OpenAI over trade secrets; OpenAI hits back publicly and seeks dismissal
Apple has sued OpenAI for allegedly stealing trade secrets and asks the court to curb OpenAI's hardware plans. OpenAI has published its emails and messages, including a misaddressed email from an…
lawyerThe case offers an example of handling trade-secret claims and litigation communications, and the misaddressed email points to lawyers' duty of care.corporate lawyerThe trade-secret risks of employee mobility and hardware development partnerships should be reassessed.tool · globalLegora reaches $5.6B valuation as Harvey reportedly seeks $15.5B
Legora has reached a $5.6 billion valuation, while Harvey is reportedly in talks to raise at least $500 million at a $15.5 billion valuation. Deals are also under way elsewhere in the market…
lawyerCompetition among legal AI tools affects law firms' choice of tools and their negotiating position on contracts.corporate lawyerAs providers consolidate, the risks of vendor dependence and data handling grow.lawsuit · United StatesAnthropic faces new lawsuit over books used for training as scanning practice comes to light
A new suit against Anthropic alleges unlawful use of published books to train AI. Unsealed court documents show the company bought, destroyed and scanned books to train Claude, while assessment of…
lawyerThe distinction between fair use and lawful acquisition is the key question in training cases.profession · globalAttorney AI rules: California, state bars, the BSB and Indonesia
California could become the first state to set AI guardrails for attorneys, and state bar associations are issuing guidance on AI use. The UK's Bar Standards Board has released comprehensive guidance…
lawyerBar rules on AI use expect transparency, supervision and risk management.corporate lawyerFirm AI policies should be aligned with the new guidance, and responsibility should also be placed on the firm.lawsuit · United StatesAI copyright suits: Adobe class action, NYT–Microsoft, Nvidia, Reddit
Authors have filed a class action against Adobe over unlicensed use of their books, while the New York Times was not allowed to add a new claim against Microsoft to its OpenAI suit. Perplexity failed…
corporate lawyerThe litigation risk facing AI developers and their suppliers is also showing up in investor communications.lawyerClass actions are coming to the fore on the plaintiff side, and jurisdictional and claim-narrowing strategies on the defendant side.law · United StatesCalifornia's AI transparency law takes effect in first phase, targeting deepfakes and scams
The first phase of California's AI transparency law has taken effect, targeting deepfakes and deception. Under Senator Becker's law, AI providers must embed identifiers in content, with fines of…
compliance specialistAI providers operating in the California market must put detection tools and disclosure obligations in place.corporate lawyerFine exposure should be built into the risk allocation in provider contracts.regulation · United StatesFederal government moves to curb state AI laws, with some states likely targets
The Trump administration wants to limit state AI laws through federal regulation, and some states could become prime targets. Following Illinois's frontier AI rules, legal advisers expect further…
compliance specialistCompliance plans for state rules will remain uncertain until federal preemption takes direction.corporate lawyerUS AI contracts need flexible change-in-regulation clauses.lawyerFederal preemption disputes will bring constitutional-law and regulatory advisory work.caseSouth African judges' rulings also cite AI-fabricated case law
A South African acting judge has been ordered to explain why a judgment contained at least 11 fabricated or inaccurate case-law citations, found at an appeal hearing. In another case, the Labour…
lawyerOn appeal, it is worth checking the citations of the opposing side and of the first-instance decision.lawsuit · United StatesMeta invokes Section 230 against fake investment ad impersonation claims
Meta is citing Section 230 in its defence against lawsuits over impersonation in fake investment ads. In the Bouck case, Judge Seeborg previously held that Section 230 does not protect the company…
lawyerFor AI-generated content, the limits of platform immunity (Section 230) and of First Amendment-based defences are central to litigation strategy.corporate lawyerPlatforms' liability risk for AI content they generate may not be covered by Section 230 protection.lawsuit · United StatesArtists' class action against Udio survives dismissal bid and moves to New York
A federal judge has declined to dismiss independent artists' copyright class action against Udio, but moved the case from Illinois to New York. UMG has separately asked for the dismissal of the US…
lawyerThe jurisdictional transfer also affects forum-selection strategy in class actions.