MLex reported that lawyers told a California state court judge in San Francisco on 24 July that the number of lawsuits alleging ChatGPT contributed to users harming themselves could rise from roughly two dozen to 50 or more. The judge began mapping a procedural course for the coordinated proceedings, which the parties expect to run for several years.
MLex reported on 27 July that the Munich Regional Court will decide the German collecting society GEMA's case against the AI music platform Suno on 31 July. The ruling will test whether AI music tools need licences to train on protected songs and whether recognisable generated outputs can infringe under German and EU copyright rules.
Transparency Coalition reported on 24 July that New Jersey legislators are advancing a cluster of narrow AI bills, including measures to bar advertising generative AI as able to practise a state-regulated profession and to regulate AI-based electronic monitoring in employment and public services. Further bills would require certain AI developers to make disclosures to the Attorney General and oblige boards of education to adopt AI policies.
Regulation
- The European Commission closed its targeted stakeholder consultation on the draft guidelines for classifying high-risk AI systems on 23 July 2026, after extending the original June deadline by four weeks at the request of trade associations. The guidelines interpret Article 6 of the AI Act across both Annex I product-embedded systems and the Annex III use cases covering biometrics, education, employment, credit, law enforcement and justice, and the Commission states that the final version will be adopted by the end of 2026.
- The European Commission sent TikTok preliminary findings on 24 July 2026 that the accounts of minors do not meet the safety standards required by the Digital Services Act. The Commission's provisional view is that minors' content should by default be visible only to accepted contacts, and that content posted by 16 and 17 year olds should not be recommended to any other user through the For You Feed.
Cases
- Via the judgment in ANI Media Pvt. Ltd. v Open AI OpCo LLC, it can be seen that Justice Amit Bansal of the High Court of Delhi dismissed ANI's application for an interim injunction on 24 July 2026, in CS(COMM) 1028/2024, reserved on 27 March 2026 and running to 135 pages. The Court held on a prima facie basis that storing ANI's original literary works to train the models underlying ChatGPT falls within Section 52(1)(a) of the Copyright Act as private or personal use including research, and so does not infringe under Section 51. It also found that the nine pleaded examples post-dated the April 2022 and April 2024 training cut-offs and so pointed to retrieval-augmented generation rather than memorisation, that the responses were not a substantial reproduction, and that the Court retained territorial jurisdiction notwithstanding that training occurred on servers outside India, while stating that its observations would not affect the final outcome of the suit.
Academia
- arXiv hosts 'Towards Agentic AI Governance: A Preliminary Assessment' by Mubarak Raji, posted 8 July 2026. The paper reviews the emerging literature on governing systems that autonomously plan and execute tasks, identifying the features that distinguish agentic AI from earlier systems and synthesising the governance priorities, mechanisms and stakeholder roles proposed so far.
- arXiv hosts 'Governing Artificial Intelligence: Public Preferences and Regulatory Options' by Magnus Lundgren and Jonas Tallberg, posted 16 July 2026. Using a conjoint survey experiment across seven countries, the authors report that citizens prioritise safety over innovation, public oversight over private self-regulation and international over national approaches, and argue this reveals a systematic misalignment with dominant regulatory designs.
Events
- InfoGov World Media will hold the AI Gov World Conference 2026 at the Flamingo Hotel in Las Vegas, with a training day and executive forum on 12 October and the conference on 13 and 14 October 2026. Hosted by the Certified Information Governance Officers Association, the programme covers governance of agentic AI, risk management and compliance across the model lifecycle, with keynotes from practitioners at Clarion AI Partners and Foley and Lardner.
- DATAVERSITY will hold the AI Governance Conference as an in-person event at the Omni Providence Hotel in Providence, Rhode Island, from 16 to 19 November 2026, co-located with the Data Governance and Information Quality Conference. Registration is open, and the organisers will also run an online edition on 30 September 2026.
Sources: MLex, Transparency Coalition, European Commission, VERDICTUM, arXiv, InfoGov World Media, DATAVERSITY