Global and Domestic Policymakers Struggle to Keep Pace with AI Advancement

Table of Contents

1. AI Safety Hits the UN Security Council — and the US Says No to Global Governance

2. Trump Hosts AI Executives at the White House, Favoring "Self-Policing" over New Regulation

3. US and China Agree to an AI "Hotline" and a Formal AI Dialogue

4. Global Policymakers Struggle to Keep Pace with AI Advancement

5. UK's Healthcare AI Blueprint

6. CMA Google Search-Choice Proposals — AI Assistants on the Menu

7. European Commission Requires Google to Open Android to Rival AI Assistants under the DMA

8. California Executive Order Fast-Tracks AI Auditor Laws and Explores an AI "Kill Switch"

9. Illinois Governor Establishes Artificial Intelligence Cabinet

10. New York Begins Implementing the RAISE Act

11. NYC Council Unveils Package of AI Regulations

12. Maryland Announces AI Framework

13. Multistate AG Coalition Puts AI and Algorithmic Decision-Making on Notice

14. South Carolina and North Dakota Issue AI Bulletins

15. New Lawsuit Highlights Growing Nuisance Risks for Data Centers

16. AI Billing Wars Driving Up Health Care Costs

This month's briefing covers the US rejection at the UN of international coordination on AI governance, as global policymakers struggle to keep pace with AI advancement. Meanwhile, the governors of California, New York, Illinois, and Maryland are advancing AI oversight, and the UK's "Healthcare AI Blueprint" recommends a regulatory framework for AI-enabled medical devices. Read on for a deeper dive into these and more key updates.

Regulatory, Legislative & Litigation Developments

AI Safety Hits the UN Security Council — and the US Says No to Global Governance

On September 23, 2026, OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, and Hugging Face CEO Clément Delangue briefed the UN Security Council on the risks of frontier AI, with Turing Award recipient and co-chair of the UN's Independent International Scientific Panel on AI, Yoshua Bengio, warning that AI beyond human control represents a threat "that none can contain alone and that does not respect the borders we defend." All three industry leaders called for international coordination on risk evaluation standards and safety disclosures, with Amodei stating that "if managed poorly... [AI] could be a risk to humanity as a whole" and pledging that Anthropic would "slow down as much as necessary" to ensure safety — but stressing that the challenge is "ultimately bigger than any one company."

The Trump administration rejected these calls outright: White House science adviser Michael Kratsios told the Security Council that the pace of AI development "is not a reason to pause" or "constrain it with new global governance structures," echoing President Trump's declaration at the General Assembly the day before that "the United States totally rejects any attempt to construct a globalist scheme to control for the artificial intelligence being spoken of so much now."

Looking ahead, the UN's Independent International Scientific Panel on AI has already warned that current safeguards "cannot keep pace" with AI's advancing capabilities, and the second session of the UN Global Dialogue on AI Governance is scheduled for May 2027 in New York — meaning the push toward a multilateral governance framework will continue even as the US charts its own course.

Trump Hosts AI Executives at the White House, Favoring "Self-Policing" over New Regulation

On September 29, 2026, President Trump and House Speaker Mike Johnson hosted leaders of major AI companies at a White House lunch. Guests included executives from Google, Meta, OpenAI, Anthropic, and other industry leaders. Afterward, the president said the executives had agreed to voluntary AI standards, including working with "independent auditors" to verify that AI systems perform as intended by their designers. He called the accord "morally binding" and said it would allow for "tremendous self-policing." He also suggested forming a 10-person committee to help oversee the industry.

The same day, he signed an executive order requiring federal agencies to use the term "super intelligence" instead of "artificial intelligence." The order does not, by itself, create a new regulatory framework. Days later, on October 4, the president named Director of National Intelligence Jay Clayton to lead a new White House "Super Intelligence Force," which will coordinate the federal government's approach to AI and has 120 days to report on the technology's risks and opportunities and what role the government should play in overseeing it.

US and China Agree to an AI "Hotline" and a Formal AI Dialogue

After Chinese President Xi Jinping's three-day state visit to Washington on September 23-25, 2026, the two countries agreed to establish a bilateral communication channel for "AI incidents" and confirmed that a formal China-US AI Dialogue is scheduled for November. Treasury Secretary Scott Bessent had proposed the incident-notification system in talks with Chinese Vice Premier He Lifeng in New York. He cited concerns ranging from cyberattacks and biological misuse to autonomous systems malfunctioning. US Trade Representative Jamieson Greer compared the channel to the Cold War "red phone" between the Kremlin and the White House. Neither government has said what counts as an incident or how the channel will work.

Global Policymakers Struggle to Keep Pace with AI Advancement

In early September 2026, the European Commission convened officials to discuss AI as a centerpiece of the bloc's agenda, with Europe's "AI czar" Jim Hagemann Snabe outlining economic opportunities while others raised safety concerns; yet the meeting produced no concrete conclusions.

European Commission President Ursula von der Leyen acknowledged in her September 16 State of the Union address that "A.I. comes with a balance of possibilities and risks," but governments remain divided on whether to ramp up regulation or risk economic disadvantage if others do not.

In the United States, congressional momentum for AI safety legislation has stalled despite bipartisan efforts; Representatives Lori Trahan and Jay Obernolte introduced a bill requiring safety testing of AI models, but it remains on ice after corporate lobbying and partisan disagreements.

At a Washington summit in late September, President Trump and President Xi Jinping stressed AI's importance to their nations but could not agree on safety protocols beyond the AI hotline, epitomizing the broader pattern of global competition trumping cooperation on AI governance.

UK's Healthcare AI Blueprint

The UK's National Commission into the Regulation of AI in Healthcare has published recommendations on 10 September 2026, advising the UK government on a regulatory framework for AI-enabled medical devices. The Commission pushes for "proportionate" regulation, covering the whole product lifecycle that works across the health system around three pillars: (1) lifecycle regulation, (2) shared system-wide responsibility and (3) trust and transparency. The recommendations include staged authorisation, continuous real-world monitoring, regulatory sandboxes, and clearer liability rules. Further, the Commission pushes for a rebalancing in favour of post-market surveillance where sufficient pre-market assessment is challenging, and for regular performance reporting for AI-enabled medical devices.

CMA Google Search-Choice Proposals — AI Assistants on the Menu

The UK Competition and Markets Authority (CMA) published proposals to give UK users of Google Android and Chrome more choice over their search service on 23 September 2026. The proposal takes into consideration the growing use of AI services such as ChatGPT and Perplexity and their effect on search habits. Under the proposals, Google would have to show a choice screen of search services when users first use their Android device or Chrome and ask again once a year. Any provider would have to attribute publisher content fairly, showing users where results came from. The proposals are open to consultation, closing October 9, with a final decision expected by year-end.

European Commission Requires Google to Open Android to Rival AI Assistants under the DMA

The European Commission issued two sets of binding measures to Google under the Digital Markets Act (DMA) on 16 July 2026. Firstly, Google must provide third-party AI services access equal to its own on the operating system of Android devices, giving users greater choice in functionality, such as using third-party AI assistants to carry out actions in apps.

Secondly, Google's data-sharing obligations to other search engines must be extended to AI chatbots offering search functions. Google will anonymize the data in several layers, following joint DMA–General Data Protection Regulation (GDPR) guidelines, and verify whether a particular recipient poses serious cybersecurity or data protection risks before sharing.

The measures also set out transparency requirements for data access and a data pricing formula. Google must start sharing search data by January 2027, and the Android changes must take place by July 2027.

California Executive Order Fast-Tracks AI Auditor Laws and Explores an AI "Kill Switch"

On September 18, 2026, California Gov. Gavin Newsom issued an executive order accelerating two AI oversight laws he signed earlier that month. The order moves the deadline for publishing criteria for certifying independent verification organizations (IVOs) to assess AI systems for safety and risk under SB 813 from January 1, 2028, to May 1, 2027; and the deadline for the state registry and standards for AI auditors under AB 1405 from January 1, 2029, to December 1, 2027.

The order also directs the state's Government Operations Agency, in consultation with other parts of state government and national experts, to submit recommendations by November 16, 2026, on whether to amend state law to:

1. Embed IVOs onsite at large frontier developers' labs to conduct periodic audits
2. Require independent verification of the safety frameworks, transparency reports, and risk assessments that frontier developers must file under state law
3. Require a "kill switch" for frontier models, with an IVO verifying on an ongoing basis that it works
4. Broaden the definition of reportable "critical safety incidents" to cover loss-of-control events

The order itself imposes no kill-switch or new substantive requirements on AI developers.

Senator Jerry McNerney plans to introduce legislation in December 2026 to implement the recommendations. Because Newsom's term ends in January 2027, any resulting bill would likely be signed or vetoed by his successor.

Any enacted law could also face a challenge from the US Department of Justice's AI Litigation Task Force, formed at the direction of President Trump's December 2025 executive order targeting state AI laws, which also directs the Commerce Department to withhold certain federal broadband funds from states with AI laws deemed "onerous."

Illinois Governor Establishes Artificial Intelligence Cabinet

On September 22, 2026, Illinois Gov. J.B. Pritzker signed an executive order establishing the Illinois Artificial Intelligence Cabinet, which will combine academic and professional experts appointed by the governor across several sectors, including technology, emergency management, commerce, and public safety and health, with representatives from state agencies. It will be tasked with developing policies and procedures to protect state assets — including buildings, water systems, education infrastructure, and information storage — from threats posed by artificial intelligence. The cabinet will also evaluate the feasibility of state regulatory measures, such as conditioning tax incentives on satisfaction of safety and transparency standards.

New York Begins Implementing the RAISE Act

New York's Responsible AI Safety and Education (RAISE) Act, signed in December 2025, requires large frontier AI developers operating in the state to publish safety and transparency frameworks, report critical safety incidents within 72 hours, and file quarterly catastrophic-risk assessments. Developers must also register with the state and pay required assessments. The New York attorney general is responsible for enforcement, with civil penalties of up to $1 million for certain first violations and $3 million for subsequent violations.

On September 21, New York Gov. Kathy Hochul announced that developer registration will begin in November, with compliance required beginning January 1, 2027. The Act will be administered by the new Office of Digital Innovation, Governance, Integrity, and Trust (DIGIT) within the Department of Financial Services. Marc Gilman has been named DIGIT's Deputy Director for the RAISE Act. Additional regulation may follow: Governor Hochul said she would explore building on the Act, including by increasing penalties, while the Act's sponsors have called for third-party audits, stronger safety frameworks, and expanded whistleblower protections.

NYC Council Unveils Package of AI Regulations

On September 25, 2026, New York City Council Speaker Julie Menin announced a package of bills to regulate AI at the city level. The centerpiece, N.Y.C. Council Introduction No. 2602, would prohibit businesses from marketing, selling, or using an AI system in the city unless a third-party validator has assessed the system for data quality, bias, privacy, and security and confirmed that it includes a human-operated "kill switch." Violations would carry a penalty of $25,000 each. Companion bills would award whistleblowers a share of fines recovered from AI companies; create a private right of action for foreseeable harms caused when third parties circumvent, or "jailbreak," safety controls; and protect city employees and contractors who report AI safety threats.

Other measures would require city contractors to report AI safety incidents within 24 hours; prohibit misleading safety claims; establish privacy rules for chatbots; and require the city to develop a response plan for AI-related incidents. The City Council held a hearing on the proposals on October 5, which included testimony from both current and former employees of Anthropic, OpenAI, and Google DeepMind. Mayor Zohran Mamdani has previously stated his view that AI regulation should be led by the federal government.

Maryland Announces AI Framework

On September 22, 2026, Maryland Gov. Wes Moore announced an AI framework outlining his administration's approach to the regulation and governance of AI. The framework, developed as a response to the lack of federal guardrails, is built on three principles: (1) protecting Marylanders, (2) centering workers, and (3) keeping children safe. As part of the framework, the administration will work toward additional regulation of frontier AI companies (including safety testing, independent third-party audits, whistleblower protections, public reporting, and 72-hour incident reporting).

The framework also includes plans to establish a statutory right of publicity to recognize an individual's likeness as a property right, develop guidance to inform citizens of current laws that reach AI-related decisions, and take action to bar algorithmic rent-setting, regulate workplace AI surveillance, and restrict AI-driven addictive design for minors.

Multistate AG Coalition Puts AI and Algorithmic Decision-Making on Notice

On September 17, 2026, a coalition of 14 state attorneys general, including California, New York, Illinois, and Massachusetts, issued joint guidance reaffirming that existing antidiscrimination laws — covering disparate impact and intentional discrimination — apply with full force to artificial intelligence and algorithmic decision-making systems used by lenders, landlords, and employers. The guidance spotlights recent enforcement wins, including a $2.5 million settlement with a student loan lender over AI-driven practices and federal court decisions sustaining disparate impact claims against tenant-screening algorithms and AI-powered hiring tools.

On the compliance front, the coalition urges public and private entities to conduct lifecycle auditing of machine learning and algorithm-based tools — from development through deployment — to identify discriminatory outcomes and refine models, warning that failure to do so invites enforcement action.

For practitioners, the takeaway is clear: the multistate coalition views algorithmic bias as a priority enforcement area, and entities relying on automated decision-making without proactive bias testing face significant and growing liability exposure.

South Carolina and North Dakota Issue AI Bulletins

On September 24, 2026, South Carolina became the 27th jurisdiction to adopt the National Association of Insurance Commissioners' model bulletin on the use of AI systems by insurers. The South Carolina bulletin follows the NAIC model in most respects, with a few notable changes. Among other things, South Carolina's bulletin is addressed to all insurers, producers, licensees, and other interested persons transacting the business of insurance in the state (and not just to insurers). In addition, the bulletin expressly states that insurers are responsible for decisions made and actions taken by their producers, adjusters, and other authorized representatives based on information derived from or generated by the AI systems. Finally, the South Carolina version includes language indicating that insurers "must comply" with the guidelines set forth in the bulletin.

On October 5, North Dakota adopted the NAIC model with some significant revisions. Among other things, the North Dakota version includes revised definitions, expects governance frameworks to cover AI use by an insurer's authorized agents and representatives, and suggests that insurers require third-party data and that model vendors maintain a written program for the use of AI systems that is commensurate with the standards applicable to insurers.

New Lawsuit Highlights Growing Nuisance Risks for Data Centers

As more data center facilities come online near residential communities, plaintiffs are increasingly bringing claims based on alleged operational impacts, particularly noise from cooling systems, generators, and other infrastructure. Most recently, residents filed a lawsuit against X.AI Corp. alleging that noise generated by infrastructure supporting its Memphis-area data center operations constitutes a nuisance and has interfered with the use and enjoyment of nearby properties. The suit is one of several recent actions that underscore the growing litigation risk facing data center developers and operators as communities increasingly scrutinize the impacts of large-scale AI infrastructure projects.

Our recent client alert, "Data Centers in the Crosshairs: How to Avoid Becoming a Target of Nuisance Lawsuits," discusses the common themes emerging and offers practical strategies that developers and operators can implement to reduce litigation risk through proactive planning, community engagement, and careful coordination of permitting and communications.

AI Billing Wars Driving Up Health Care Costs

Recent news reports indicate that hospitals and health insurers are both using AI in their long-running fight over medical billing, and that this is pushing health care costs higher. A September 2026 Blue Cross Blue Shield Association analysis estimates that hospitals' AI-assisted coding added about $942 million in health care spending between 2023 and 2025, leading to higher premiums and out-of-pocket costs for families, employers, and taxpayers. The analysis attributes much of the increase to added secondary diagnoses, such as anemia or low sodium, that moved more than 55,000 inpatient stays into higher-paying categories with no change in treatment.

Insurers face criticism for using AI to delay or deny care, including in prior authorization and Medicare's WISeR (Wasteful and Inappropriate Service Reduction) pilot, which lawmakers and patient advocates say has caused inappropriate denials and longer waits. According to these reports, patients are caught in the middle with unpaid bills or denied care, and benefits advisers cite AI as one reason employer health costs could rise as much as 11% next year.