In a NYC council hearing, representatives from Anthropic, OpenAI, Google, and Meta testified about AI risks. Former DeepMind researcher Alex Turner warned misaligned AI could be more dangerous than China, and Anthropic whistleblower Jacob Coxon called the industry’s approach “extremely reckless.” Council Speaker Menin expressed frustration that no company could quantify catastrophic risk.
What Happened
Former Google DeepMind researcher Alex Turner warned that “misaligned” artificial intelligence could prove more dangerous than China’s aggressive development push during a New York City Council hearing on Monday. “China is not our only potential adversary,” he said. “With reasonably high chance, we are racing to build and grow our own adversary here at home, which is misaligned AI. Misaligned AI is everyone’s adversary, including our own, and one day may be more powerful than China.”
The hearing was prompted by a question from Council Speaker Julie Menin on how to balance frontier model development with the technology race with China. Anthropic whistleblower Jacob Coxon, who raised early concerns about the dire risks of artificial intelligence, called out the industry’s “extremely reckless” approach. “On the current path, I think it is more likely than not that humanity loses control to these AIs, and it could end in human extinction,” he told a packed room.
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
— Jacob Coxon (@hilbertspaess) September 9, 2026
Senior leaders from Anthropic, OpenAI, Google, and Meta testified under oath about the potential dangers of the technology and outlined current safety measures. The four companies sent representatives: Anthropic’s Logan Graham, head of its Frontier Red Team; OpenAI’s Morgan Dwyer, head of policy development and operations; Google’s Alice Friend, director of AI and emerging tech policy; and Meta’s Shane Cahill, AI policy director for legislation.
Menin appeared frustrated at times with the lack of direct responses, including one on legal liability when AI models cause serious financial harm or even death. “I’m going to take it then that neither of the four of you, no company here, can quantify the risk of something cataclysmic happening,” she said, calling the answers “troubling at best.”
Google’s Friend spoke in support of a “comprehensive” federal framework to address mounting AI risks, citing national security concerns.
What This Means For You
If you work in a regulated industry—finance, healthcare, or energy—this hearing signals that federal oversight of AI is moving from discussion to concrete action. Expect new compliance requirements that may mandate risk assessments, impact statements, and third‑party audits for any deployment of large language models or generative AI. Prepare by documenting your model development lifecycle, establishing internal safety protocols, and staying abreast of evolving regulations.
For product managers and engineers, the message is clear: “misaligned” AI is a real threat. Integrate alignment checks into every stage of model training. Use red‑team testing, as Anthropic’s Frontier Red Team does, to surface unintended behaviors before release. Consider adopting open‑source alignment frameworks or collaborating with academia to benchmark safety metrics.
If you’re a startup founder, the hearing underscores the importance of building safety into your value proposition. Investors are increasingly scrutinizing AI safety practices. Demonstrating a robust safety framework can differentiate you in a crowded market and mitigate future regulatory burdens.
For policymakers, the testimony highlights a gap: the industry cannot quantify catastrophic risk. This uncertainty should drive the development of probabilistic risk models and scenario planning tools. Governments may also need to legislate liability standards that hold companies accountable for AI‑induced harm, especially in high‑stakes sectors.
In the near term, watch for the release of a federal AI safety framework. The hearing’s emphasis on a “comprehensive” approach suggests that Congress may draft legislation that includes licensing, testing, and reporting requirements. Keep an eye on the NYC Council’s follow‑up meetings, as local regulations often set precedents for national policy.
Why It Matters
This echoes concerns raised in the recent AI in Health Inspection & Compliance piece, where industry leaders warned that unchecked AI could compromise patient safety. Both stories underline a common theme: without rigorous safety protocols, AI’s benefits may be outweighed by catastrophic risks.
The hearing also signals a shift in the U.S.–China technology race. While China’s rapid AI development has been a focal point, Turner’s warning that misaligned AI could outpace any single nation’s capabilities reframes the competition as a global existential issue. If the U.S. fails to regulate its own AI ecosystem, it risks becoming the very adversary it fears.
Moreover, the lack of quantifiable risk assessment, as Menin pointed out, could erode public trust. Transparency in risk evaluation is essential for maintaining societal confidence in AI technologies. The industry’s current “reckless” trajectory may accelerate calls for stricter oversight, potentially stifling innovation if not balanced with practical safeguards.
Key Takeaway
- Federal AI safety legislation is likely to follow this hearing, demanding rigorous risk assessments and reporting.
- Companies must embed alignment and red‑team testing into every development cycle to mitigate misalignment risks.
- Investors and regulators will increasingly scrutinize AI safety practices; demonstrating compliance can be a competitive advantage.
- The U.S. faces a dual threat: external competition from China and internal risks from misaligned AI that could surpass any single nation’s power.
Leave a Reply