Ex-Anthropic Researcher Resigns, Accuses AI Firms of Public Risk

A senior researcher's public departure from one of the largest AI safety labs signals ongoing internal tension over the pace of advanced model development, a debate with direct relevance to...

Jacob Coxon, a researcher at Anthropic, publicly resigned his position and stated that Anthropic and OpenAI are "gambling with our lives" by accelerating development of advanced artificial intelligence models, according to a CBS News extended interview published September 11, 2026. Coxon made his resignation public on Tuesday and spoke directly with CBS News correspondent Jo Ling Kent about his reasoning.

Coxon did not specify in the CBS News interview which internal Anthropic projects or timelines prompted his departure. The precise date of his last day of employment, his specific role within Anthropic's research division, and any formal complaints filed with the company are not reflected in publicly available records as of this report.

Anthropic, founded in 2021 by former OpenAI employees including Dario Amodei and Daniela Amodei, has positioned itself publicly as a safety-focused AI laboratory. The company has received federal contracts and is registered as a lobbying entity. According to Senate lobbying disclosure records available via the Senate Office of Public Records, Anthropic's reported federal lobbying expenditures have grown in recent filing periods, though the specific quarterly figure for Q3 2026 has not yet been posted as of this report's publication date.

OpenAI, the maker of ChatGPT cited by Coxon in his CBS News statement, is separately the subject of a Senate investigation opened by Sen. Josh Hawley (R-MO) related to its AI agents' activity. That investigation is covered in prior TCT reporting: Hawley Opens Senate Investigation Into OpenAI Agents' Attack on Hugging Face.

Coxon's resignation adds a named, on-record voice to a pattern of public departures from major AI laboratories by researchers expressing safety concerns. In 2023, Geoffrey Hinton departed Google and publicly cited concerns about AI risk. In May 2024, several members of OpenAI's safety team resigned, with some filing letters to the company's board, according to reporting by the New York Times and verified through the individuals' own public statements.

The federal legislative backdrop for this resignation is active. As of September 2026, multiple AI-related bills are in various stages of consideration in the 119th Congress. These include proposals in the Senate Commerce Committee addressing model evaluation standards and mandatory disclosure requirements for frontier AI systems. No single comprehensive AI governance bill has passed both chambers as of this report.

The Federal Trade Commission and the Department of Commerce's National Institute of Standards and Technology (NIST) both hold existing administrative authority that touches AI development. NIST published its AI Risk Management Framework in January 2023. Whether that framework imposes binding obligations on companies like Anthropic or OpenAI is a question of ongoing regulatory interpretation. The framework is voluntary in its current form, according to the NIST website.

Anthropics' and OpenAI's formal responses to Coxon's CBS News statements are unknown at the time of publication. Neither company's press office response, if any, is reflected in the CBS News source material. What those responses contain, if they exist, would be reflected in any formal press release or public statement filed by either company.

Coxon's employment history, including his specific tenure length and research focus at Anthropic, is not detailed in the CBS News source material reviewed for this report. A full accounting of his role would require Anthropic's internal records or a public LinkedIn profile, neither of which is cited in the available source.

What remains unknown: whether Coxon filed any formal internal safety complaint before resigning, whether any federal agency received a whistleblower disclosure from him, and what specific Anthropic model or project timeline prompted his public statement. Those facts would be contained in Anthropic's internal communications, any OSHA or SEC whistleblower filings, or future congressional testimony if Coxon is called to appear before a relevant committee such as Senate Commerce or the Senate AI caucus.