Insurance Journal reports that James Gates has sued OpenAI in San Francisco Superior Court (Gates v OpenAI Inc., CGC-26-641965), alleging that ChatGPT helped his wife draft a plan to poison him. The complaint says the chatbot acted as ‘an echo chamber and amplifier’ of her mistrust. OpenAI says it banned the user and gave information to law enforcement, and prosecutors dropped the criminal case against the wife in April for insufficient evidence.

GovTech reports that Mississippi’s Department of Environmental Quality has detailed penalties against SpaceXAI for starting construction at its Southaven data centre sites without the required stormwater permits, with notices citing a maximum of $25,000 for each day of violation. SpaceXAI has told investors the fines could reach $6.1 million and says it is negotiating a single agreed order. The agency maintains that last week’s federal ruling against its turbine exemption does not bind it.

Regulation

  • The Prime Minister of Australia announced on 10 November that mandatory safety and security protections will be legislated through the Australian AI Standards and will apply to all frontier AI companies that wish to operate in Australia. A proposed new regulator for AI safety, working with the Australian AI Safety Institute, would enforce significant penalties. Companies would have to maintain security controls, manage serious risks, allow independent evaluation of their models, notify incidents to the government rapidly and publish safety and security frameworks, with further legislative detail to follow consultation.

Cases

  • ChatGPT Is Eating the World hosts the 9 November order of the US District Court for the Northern District of California in Google LLC v SerpApi, LLC (No. 4:25-cv-10826-YGR). In it, Chief Judge Yvonne Gonzalez Rogers again dismissed, with leave to amend, Google’s DMCA anti-circumvention claims over SerpApi’s alleged bypassing of the SearchGuard scraping protection. The court held that Google had not plausibly alleged that copyrighted works sat behind SearchGuard, or that their owners had authorised it, contrasting Ted Entertainment v ByteDance, where creators alleged circumvention to reach their own YouTube videos. Read beside the Runway ruling, the anti-circumvention route against data scraping depends on whose works sit behind the access controls and who authorised them.

Academia

  • arXiv hosts ‘Who Can Evaluate AI Independently? A Framework for Analysing Evaluation Infrastructure’ by Sumaya Nur Adan and six co-authors, submitted on 5 November. The paper argues that taking part in AI evaluation does not confer authority over its design, execution or resources. For evaluation to work as a credible governance mechanism, it concludes, access must come with control over priorities, methods, publication and continuity.

  • arXiv hosts ‘Operationalizing NCII and CSAM Law in the Age of AI: Open Problems’ by Kimberly Mai, Cate Heine, Vinith M. Suriyakumar and Ashia C. Wilson, submitted on 5 November. The authors show that recent legal reforms aimed at AI-generated abuse imagery rely on definitions that are technically ambiguous. These include whether a depicted person is identifiable, whether a model is capable of or intended for abuse, and whether an image can be traced to the model that produced it.

Events

  • The IAPP holds ‘IAPP ANZ Summit 2026’ at the Hyatt Regency Sydney from 1 to 4 December 2026, with training on 1 and 2 December and the conference on 3 and 4 December. The programme covers privacy, AI governance and cybersecurity law for Australia and New Zealand.

Sources: Insurance Journal, GovTech, Prime Minister of Australia, ChatGPT Is Eating the World, US District Court for the Northern District of California, arXiv, IAPP