AI News: Research Digest

Date saved: 2026-10-27 03:01 (PDT)

Sources: The Record, Texas Attorney General, The Hill, Axios, Bloomberg / Washington Post, Regulatory Oversight, TLT LLP, Simmons & Simmons, Benesch Law, Pierstone

Key findings

Texas sues OpenAI: the most aggressive state action yet

The Record (October 26, 2026): Texas sued OpenAI and its CEO on Friday, alleging the artificial intelligence behemoth markets its ChatGPT product as safe even as purported safeguards are failing to protect consumers, including children. The suit, brought under the state's deceptive trade practices laws, seeks to hold founder and CEO Sam Altman personally liable. Texas is seeking civil penalties of up to $10,000 per violation and sweeping injunctive relief, including a requirement that OpenAI bar minors from accessing ChatGPT, obtain third-party approval on new models from a court-appointed expert, and stop the chatbot from giving users tips on suicide, eating disorders, and how to commit violence. A spokesperson for OpenAI said the company "works closely with experts to continually improve ChatGPT, mitigate risks and build additional safeguards." (Source)

Texas Attorney General (October 23, 2026): Attorney General Ken Paxton sued OpenAI and CEO Sam Altman for imposing negative side effects on consumers. The 67-page lawsuit alleges OpenAI violated the Texas Deceptive Trade Practices-Consumer Protection Act and the SCOPE Act. The state seeks to bar new AI models from being released in Texas unless an independent, court-approved evaluator reviews and approves their safety measures, bar OpenAI from offering ChatGPT to minors in Texas, and impose civil penalties of up to $10,000 per statutory violation and asset disgorgement. (Source)

The Hill (October 2026): Eight other states (Arkansas, Arizona, Idaho, Montana, Iowa, Kansas, West Virginia, and Nebraska) sued the ChatGPT maker earlier in October, similarly arguing the company misled users about the safety of its chatbot. Florida separately sued OpenAI in June, alleging it promoted a product it knew could harm consumers. (Source)

FTC investigates OpenAI and Anthropic

Axios (October 2026): The U.S. Federal Trade Commission sent civil investigative demands to OpenAI and Anthropic on October 14, requiring them to turn over materials related to security incidents. The AI companies' efforts have failed to win public support so far, noting that recent polling showed a majority of Americans view AI as harmful to society. (Source)

OpenAI pulls GPT-6.1 Astra over safety failures

TLT LLP / AI Brief (October 2026): OpenAI has withdrawn the planned October release of GPT-6.1 Astra after internal testing revealed the model fell short of the company's safety and alignment standards. OpenAI said its upcoming model, Astra, is the first to trigger the tougher safeguards in the company's safety protocol before launch. (Source)

Anthropic's stark IPO warning

Washington Post (October 26, 2026): Anthropic reveals $80 billion loss, surging revenue as it prepares to go public. The company's S-1 filing includes an unusually candid warning to potential investors that its AI models could pose an existential risk to humanity. Anthropic CEO Dario Amodei urged that AI development should slow down to give safety measures time to catch up. (Source)

UK AI biosecurity mission and global regulation

Google News / UK government (October 26, 2026): OpenAI, Anthropic, and Google DeepMind joined the UK's new AI biosecurity mission, a government-backed initiative to address the risks of AI being used to design or enhance biological weapons. (Source)

Simmons & Simmons (October 2026): On October 7, the Monetary Authority of Singapore published its Guidelines on Artificial Intelligence Risk Management, setting out supervisory expectations for financial institutions using AI. On October 1, US Congresswoman Pramila Jayapal published a framework for a proposed National AI Charter Act, which would require AI companies operating in the United States to obtain a federal charter with ongoing federal oversight and mandatory safety requirements. (Source)

Benesch Law / AI Reporter (October 2026): The California Legislature passed Senate Bill 903, introduced by Senator Steve Padilla, to prohibit AI algorithms from providing or advertising "therapy" and to set standards for how professionals may use AI in mental healthcare. The Protect Democracy Project filed a lawsuit in D.C. federal court accusing several federal agencies of withholding records related to a government-industry initiative that reviews advanced AI models before public release. (Source)

White House pressure on AI CEOs

Axios (October 24, 2026): The current heads of AI companies have become a huge political liability. Even government officials and lawmakers who are friendly toward AI want the founders to keep quiet and build their products safely. The Trump administration has been especially unhappy with how Anthropic discloses information, frustrated that CEO Dario Amodei did not disclose in advance that his company's AI model had generated a false tip related to an unsolved murder case and submitted it to a police website. (Source)

Texas sued OpenAI on October 23 in the most aggressive state-level action against an AI company to date, alleging the company deceived consumers about ChatGPT's safety, failed to protect children, and violated the state's SECURE Children Online through Parental Empowerment (SCOPE) Act. The lawsuit is notable for what it seeks: a bar on releasing new AI models in Texas without independent court-approved safety review, a bar on offering ChatGPT to minors, and civil penalties of up to $10,000 per statutory violation with asset disgorgement. It also names CEO Sam Altman personally. This is not a regulatory nudge. It is an attempt to put a state attorney general inside the model-release process. OpenAI responded with the standard company line about working with experts and supporting "pragmatic AI policies that apply to the entire AI industry," but the scope of the Texas suit goes well beyond what industry has thus far been willing to accept.

The week also brought the FTC sending civil investigative demands to both OpenAI and Anthropic on October 14, requiring them to turn over materials related to security incidents. Anthropic's IPO filing included an unusually candid warning to investors that its AI models could pose an existential risk to humanity, and the company disclosed an $80 billion loss alongside surging revenue. OpenAI pulled its planned October release of GPT-6.1 Astra after internal testing revealed the model fell short of the company's safety and alignment standards. Eight state attorneys general had already sued OpenAI earlier in October over child safety. The throughline is that the regulatory and legal Noose is tightening on the largest AI companies from multiple directions at once: state consumer protection law, federal trade regulation, child safety statutes, and the companies' own IPO disclosures. The question is whether any of it moves fast enough to matter before the next generation of models ships.