OpenAI and Anthropic Are Facing Pressure From Within: Researchers Want a Bigger Say on AI Safety

A small but powerful group of elite artificial intelligence researchers is emerging as an unexpected force inside the companies leading the AI race, challenging executives over safety and regulation while complicating Silicon Valley's relationship with Washington, a new report detailed.

According to Axios, researchers at OpenAI, Anthropic and other major AI companies are wielding unusual influence over corporate policy, with some publicly criticizing decisions made by their employers and pushing executives toward stronger safety measures and government oversight.

An intense recruiting battle for AI talent last year pushed compensation packages for some of the industry's most sought-after researchers into the hundreds of millions of dollars. Researchers can also receive favorable vesting arrangements and considerable freedom to discuss their work, competitors, and the broader AI race publicly.

That influence is now extending beyond laboratories and into corporate political strategy. Axios pointed to OpenAI President Greg Brockman's decision to abandon plans for an additional $25 million donation to a pro-AI political group that had faced criticism from some researchers.

OpenAI researchers and employees also say internal pressure helped move the company away from resisting certain regulations and political candidates who supported them and toward greater cooperation with lawmakers in several states.

OpenAI recently supported a transparency measure in Illinois and, late in the legislative process, backed an independent auditing bill in California. "We have a lot of debate inside OpenAI, and that's a good thing, but what ultimately drives our policy positions is what we're seeing in our own technology and how quickly its capabilities are advancing," an OpenAI spokesperson told Axios.

However, OpenAI said Thursday that it had fired three employees for allegedly mishandling sensitive information, saying their actions violated the "deep level of trust" required to keep its systems secure.

Some policymakers view the industry's regulatory positions as inconsistent because companies have opposed certain proposals before later supporting others. One source close to the Trump administration told Axios that OpenAI was "no longer seen as or acting as a leader in driving policy," claiming researchers had "neutered" the company's ability to set its own course.

A White House official offered a sharply different assessment, describing OpenAI as an important partner and pointing to a recent voluntary AI safety agreement signed with President Donald Trump.

Researchers reportedly complicated Anthropic's efforts to reach an agreement with the Pentagon following a confrontation between the company and senior administration officials earlier this year.

Employees from OpenAI, Anthropic and Google have also joined public campaigns to influence government and industry policy. More than 1,000 workers signed the "Pacing the Frontier" petition, while Anthropic researcher Jacob Coxon publicly criticized the company's safety practices after resigning. Several Anthropic employees subsequently endorsed his concerns.