Former Anthropic Researcher Jacob Coxon to Testify at NYC AI Hearing as Lawmakers Weigh AI Safety Rules
Former Anthropic researcher Jacob Coxon is set to testify before the New York City Council as lawmakers consider new rules aimed at controlling the risks associated with increasingly powerful artificial intelligence systems.
Coxon, who left Anthropic last month after raising serious concerns about the pace of AI development, will appear at the request of City Council Speaker Julie Menin. The hearing is bringing AI safety critics together with representatives from major technology companies including Anthropic, OpenAI, Google and Meta.
The hearing comes at a time when the AI industry is facing a more intense debate over whether rapid technological progress is moving faster than the safeguards designed to control it.
For investors, the discussion matters because tougher AI regulation could eventually affect how companies develop, deploy and monetize advanced AI systems.
Why Jacob Coxon’s Testimony Matters
Coxon's appearance is notable because he is not an outside critic who has never worked inside the industry.
He previously conducted AI research at Anthropic and has publicly warned about the potential consequences of developing increasingly capable systems without adequate safeguards. After leaving the company, he accused Anthropic and OpenAI of taking excessive risks in their race to develop more advanced AI.
Coxon has argued that some people working on advanced AI believe the technology could eventually become dangerous enough to threaten humanity.
Those statements are his warnings and assessments, rather than established predictions. They nevertheless have gained attention because they come from someone with experience inside a leading AI laboratory.
His testimony could therefore give New York lawmakers a perspective on how AI safety concerns are viewed from inside the industry.
New York City Wants a Larger Role in AI Regulation
The New York City Council has been preparing a package of proposals intended to strengthen safeguards around AI.
The Council announced the proposals in September and scheduled a rare Committee of the Whole hearing for October 5, bringing all 51 council members together to examine the potential risks posed by rapidly developing AI technology.
The effort is significant because much of AI regulation in the United States is still being shaped at the federal and state levels.
New York City is attempting to establish rules that could directly affect businesses operating or deploying AI systems within the city.
That creates an important question for technology companies: how much regulatory responsibility should be imposed at the local level when AI products are often developed and deployed across multiple jurisdictions?
Proposed AI Rules Include Third-Party Validation
One of the most consequential proposals would require certain AI systems to undergo third-party validation before they could be marketed, sold or deployed in New York City.
Third-party validation essentially means an independent organization would assess an AI system rather than allowing the developer to determine by itself whether the system is safe and ready for deployment.
The proposed framework would cover areas including task performance, data provenance, bias, privacy, security and safety. It would also require a technical capability for a human operator to shut down the system.
The proposal includes a potential civil penalty of $25,000 per instance for marketing, selling or deploying an AI model without the required validation or shutdown capability. Validators could also face penalties for falsifying validation.
For AI developers, such requirements could add another layer of compliance before products reach customers.
Whistleblowers Could Receive Financial Incentives
Another proposal could change the incentives for employees who report potential violations.
The Council has proposed allowing AI whistleblowers to receive a portion of fines or penalties recovered from companies following violations reported by those individuals. The Council describes the measure as a way to encourage insiders to report serious misconduct or regulatory breaches.
That could have wider implications for the AI industry.
AI companies depend heavily on researchers and engineers who understand the internal behaviour of complex models. If employees believe they can report serious safety failures while receiving legal or financial protection, more information about internal AI incidents could reach regulators.
For companies, however, this could also increase compliance and litigation risks.
AI Companies Are Being Asked to Testify Under Oath
The hearing is unusual not only because of Coxon's participation but also because major AI companies are being asked to publicly address lawmakers.
The Council said Anthropic, OpenAI, Google and Meta agreed to participate after the threat of subpoenas, while Meta had committed earlier. The Council also subpoenaed Elon Musk's AI unit after it failed to respond to requests to appear.
Representatives from the companies are expected to address AI safety and potential regulatory solutions.
That places company executives and policy officials in an increasingly important position: they must explain how their organisations evaluate risks while simultaneously convincing investors and customers that rapid AI development remains commercially viable.
Why AI Regulation Could Matter to Investors
AI regulation is no longer simply a technology-policy issue.
The industry has attracted enormous amounts of capital, and valuations of companies connected to AI depend partly on expectations that the technology can scale rapidly.
Additional regulation could affect that growth in several ways.
Higher compliance costs: Companies may need additional teams, audits, testing and documentation before deploying AI systems.
Slower product launches: Mandatory validation could increase the time between model development and commercial deployment.
Higher liability exposure: Rules allowing individuals to sue AI companies for certain foreseeable harms could increase potential legal costs.
Greater demand for safety technology: On the other hand, regulation could create new markets for AI auditing, model evaluation, cybersecurity, monitoring and compliance software.
This means regulation will not necessarily be negative for the entire technology ecosystem. Some companies could benefit if demand shifts toward independent AI testing and safety infrastructure.
A Separate Proposal Could Expand Legal Liability
New York lawmakers are also considering legislation that would allow individuals to bring claims against AI companies for certain foreseeable harms arising from malicious use or circumvention of safety controls.
Under the Council's proposal, a company could potentially face liability where the harm was foreseeable, reasonable safeguards were not implemented and a third party exploited that failure.
The concept could become particularly important as AI systems move from generating text and images toward operating as agents capable of taking actions.
An AI agent may be able to interact with websites, software, databases or other digital systems with less direct human intervention. That increases its usefulness, but it can also make mistakes or create security problems at a larger scale.
The regulatory challenge is determining where responsibility should sit when an AI system causes harm: with the developer, the user, the company deploying it or another party that exploited the system.
The Bigger Debate: Speed Versus Safety
Coxon's testimony comes amid a broader disagreement inside the AI industry.
Some technology leaders argue that society should continue developing increasingly capable AI while improving safety mechanisms alongside it. Others believe the pace of development itself needs to slow until researchers can demonstrate stronger control over frontier systems.
OpenAI CEO Sam Altman recently argued that the potential benefits of AI justify accepting some risks, while emphasizing the importance of keeping the technology broadly accessible. Anthropic has taken a more cautious public position on the pace of frontier development.
That difference is becoming increasingly relevant for investors.
If regulators impose stringent controls, companies may face slower growth but potentially lower long-term safety and liability risks. If regulation remains relatively light, AI development could move faster, potentially increasing both economic benefits and the consequences of serious failures.
What Investors Should Watch Next
The New York hearing is unlikely to settle the national AI regulatory debate, but it could provide important signals.
Investors should watch:
Whether New York advances third-party AI validation requirements
How lawmakers define a mandatory human shutdown capability
Potential rules governing AI whistleblowers
Liability standards for AI-related harm
Responses from Anthropic, OpenAI, Google and Meta
Whether other U.S. cities or states adopt similar rules
The effect of regulation on AI development costs
Growth in AI safety, auditing and compliance businesses
The market impact will depend less on one city hearing than on whether similar policies spread to other major technology centres.
Conclusion
Jacob Coxon's planned testimony puts an industry insider and AI safety critic directly before New York lawmakers as the city considers some of the most ambitious local AI safeguards in the United States.
The proposed rules could require independent validation, human shutdown capabilities, stronger whistleblower incentives and new avenues for legal action. If adopted, they could raise compliance costs for AI developers while creating opportunities for companies specialising in safety and oversight.
For investors, the key issue is not simply whether AI regulation becomes stricter. It is whether regulation can slow dangerous development without undermining the economic value being created by the technology.
The October 5 hearing could provide an early indication of where that balance is heading.
Follow our blog for more AI business news, technology developments, global markets and investment analysis.
This article is for informational and educational purposes only and should not be considered investment advice.