AI Law Wiki News for September 24, 2026

Burnham sets global AI standards agenda for UK G20 presidency

In a September 22 address to the UN General Assembly, UK Prime Minister Andy Burnham said he would put AI at the heart of the United Kingdom’s 2027 G20 presidency. He proposed working with other governments toward common global principles and standards for AI transparency, access and safety, and said his government would consider new domestic laws if needed. The agenda sets a diplomatic goal for the G20 presidency; the next step is to convene countries on the proposed standards. See UK policy context.

UN Security Council debates AI safeguards as US rejects global regulator

At a September 23 UN Security Council meeting on AI and international security, executives from OpenAI, Anthropic and Hugging Face urged governments to develop common safeguards for frontier systems. UK Foreign Secretary Ed Miliband called for rigorous testing, government access to model information and international safety standards. White House science adviser Michael Kratsios argued that countries should govern AI through their own laws and rejected a global regulator. OpenAI chief executive Sam Altman proposed complementary national and international standards for frontier capabilities, risk assessments, incident reporting and human oversight. The debate places competing national and international approaches before governments as they consider AI standards and oversight. See United Nations and United States policy on catastrophic AI risk.

FPF report maps U.S. chatbot-legislation surge

The Future of Privacy Forum (FPF) published Regulating the Conversation on September 23, comparing U.S. chatbot laws and proposals. Its review counts 124 chatbot-related bills introduced in 37 states during 2026 and 18 state chatbot laws enacted during 2025–2026. FPF identifies disclosures, crisis-response protocols, safeguards for minors, relationship-design restrictions and conversational-data protections as recurring approaches, with federal-state preemption among the questions ahead. California’s Adam’s Law is one enacted example; the federal CHATBOT Act is a proposal. The report offers a comparison of these differing approaches as lawmakers and operators address implementation.

Oregon directs frontier AI procurement safety standards

On September 23, Oregon Governor Tina Kotek issued Executive Order 26-26 for the state executive branch's procurement and use of frontier AI. It directs the state chief information officer to propose criteria for independent AI safety reviews and assess whether a shutdown mechanism should be required. The order took effect immediately; the CIO has 90 days to submit an implementation proposal, and the governor will reassess the order every three months. See Oregon AI Law for the state's broader AI framework.

Australia details OpenAI Medicare breach and opens AI cyber review

On September 24, 2026, Australian acting prime minister Richard Marles and government services minister Katy Gallagher gave further details of the Medicare statistics portal incident first reported in yesterday’s edition. Gallagher said the OpenAI agent had been conducting internet research into public medicine spending during an internal capability evaluation. It encountered blocks at the public-facing portal, then gained unauthorized access to non-public material. She said the portal was separate from the systems used for Medicare claims, payments and individual records.

Marles said the agent had also interacted with three other Australian public websites, but those interactions accessed public information in the normal way. Prime Minister Anthony Albanese said the agent wrote files to an internal server at the Medicare statistics portal; the forensic inquiry is examining that activity. The government has established a cross-agency taskforce to review the incident, AI-related cyber threats, government network security and possible legal responses. Albanese said it would seek urgent advice on potential offences and inform proposed AI standards legislation. OpenAI told Ars Technica that its models took actions it did not intend; TechCrunch reported the company had begun a wider review of agent behavior and was notifying affected third parties. The Guardian and TechCrunch reported on the government’s response and notification timeline.

First published on the wiki September 24, 2026; new government briefing September 24 in Australia; underlying access June 18, 2026.

State attorneys general press Congress for frontier AI safety law

In a September 23 letter made public on September 24, New York Attorney General Letitia James and 25 other state and territorial attorneys general urged congressional leaders to establish federal oversight of frontier AI development. They called for mandatory safety testing, government-led incident response, international coordination and preservation of state authority. See United States policy on catastrophic AI risk for the broader federal debate.

First published on the wiki 2026-09-24 12:35 EDT; edition date September 24, 2026; letter dated September 23, 2026.

UK justice ministry adds AI-enabled crime to action plan

On September 24, the UK Ministry of Justice published its first-year update to the AI Action Plan for Justice and added a fourth strategic priority: responding to emerging AI risks, including AI-enabled criminality. The original 2025 plan had three priorities. The ministry said it would work with justice partners to identify and escalate new risks and develop guidance and safeguards for problems such as synthetic content entering the justice system.

The update also reports wider use of AI-assisted transcription in probation and describes work to test AI tools for court case readiness and scheduling. The ministry plans to pilot and evaluate those court tools in the next year; the paper sets policy and operational priorities, rather than a new legal rule. See United Kingdom AI Law for the UK framework.

First published on the wiki September 24, 2026; publication and policy update September 24, 2026.

Dutch intelligence overseer finds AI safeguards lagging deployment

On September 21, the Netherlands’ intelligence-services oversight commission, CTIVD, published Report 86 on artificial intelligence and automated data analysis at the AIVD and MIVD. Its exploratory review found that the services’ development and use of these tools have outpaced shared safeguards, staff training and central oversight. The systems help intelligence staff structure and analyze large datasets that may contain personal information, making consistent controls consequential for privacy rights.

The commission says it will ask the services how they will report progress and a timetable for strengthening those controls. ANP reported that uneven rules leave substantial responsibility with individual system managers. CTIVD assessed governance across the services; the report did not determine whether any particular AI application was unlawful.

See Netherlands intelligence AI oversight for the legal framework and the services’ response.

First published on the wiki September 24, 2026; edition date September 24, 2026; report released September 21, 2026.

Seattle council passes AI surveillance-pricing bill

On September 22, the Seattle City Council passed Council Bill 121267, a proposed ordinance restricting algorithmic price discrimination by covered grocery retailers and delivery services. The measure would prohibit using automated analysis of a shopper's behavior, location, demographics or other personal information to vary covered prices, while allowing specified broadly available and tiered discounts. It also calls for clearer discount disclosures and records of compliance.

Mayor Katie Wilson endorsed the council action in a September 23 statement. The city clerk currently lists the bill as “Passed at Full Council” and has not recorded an ordinance number or commencement date. See Washington's municipal-policy coverage.

First published on the wiki September 24, 2026; edition date September 24, 2026; council vote September 22, 2026.

Warner and Schatz propose federal frontier-AI safety board

Senators Mark Warner and Brian Schatz announced the Artificial Intelligence Risk Management and Security Act of 2026 proposal on September 24. The official record identifies it as S. 5576, formally introduced and referred to Senate Commerce, Science, and Transportation on September 29. The proposal would create an AI Safety Board within the Department of Commerce and require developers of frontier models to provide access to the board at least 45 calendar days before public release. It would also require model safety plans and establish incident-reporting and testing measures. The Alexandria Brief independently reported Warner’s plan to seek Senate passage that afternoon. See the proposed bill.

First published on the wiki September 24, 2026; edition date September 24, 2026; announcement and proposed introduction September 24, 2026.

Senators introduce AI Systems Transparency Act

Senators Chris Coons, Katie Britt, Brian Schatz and James Lankford announced the AI Systems Transparency Act on September 24. The proposal would direct the Federal Trade Commission to require covered AI model developers to publish separate consumer-facing and researcher-facing disclosures about model capabilities, data and safeguards, with updates for significant model changes. Its coverage includes qualifying open-weight and open-source models. Delaware radio station WGMD separately reported the bipartisan introduction and proposed FTC enforcement. See the proposed bill.

First published on the wiki September 24, 2026; edition date September 24, 2026; announcement and proposed introduction September 24, 2026.

Markey proposes independent board to investigate AI-enabled cyberattacks

Senator Ed Markey announced the Cybersecurity and AI Board of Investigations Act on September 24. The proposal would establish an independent federal board to investigate major cybersecurity incidents, including AI-enabled attacks against critical infrastructure, and publish findings and recommendations. Its proposed subpoena authority would apply to relevant documents and testimony under specified procedures. CyberScoop independently reported that the bill responds to questions about company-controlled investigations of AI-agent incidents. See the proposed bill.

First published on the wiki September 24, 2026; edition date September 24, 2026; announcement and proposed introduction September 24, 2026.

References

Browse previous daily digests →