AI safety researcher quits Anthropic, warning 'world is in peril'

A pair of researchers resigned from Anthropic and OpenAI this week, with one warning that the “world is in peril” from a multitude of crises.

Mrinank Sharma announced his resignation from Anthropic in an open letter to his colleagues Monday. Sharma, who has served on the company’s technical staff since 2023, first noted that he “achieved what I wanted to here” and is “especially proud of my recent efforts to help us live our values via internal transparency mechanisms; and also my final project on understanding how AI assistants could make us less human or distort our humanity.”

“Nevertheless, it is clear to me that the time has come to move on,” he wrote later in the letter. “I continuously find myself reckoning with our situation. The world is in peril. And not just from AI, or bioweapons, but from a whole series of interconnected crises unfolding in this very moment.

“[Throughout] my time here, I’ve repeatedly seen how hard it is truly let our values govern actions,” Sharma added. “I’ve seen this within myself, within the organization, where we constantly face pressures to set aside what matters most, and throughout broader society too.”

Two days later, Zoë Hitzig wrote in an op-ed for The New York Times that she was resigning from OpenAI over the company deciding to test ads on its ChatGPT platform. Hitzig, a research scientist with the company since 2024, said that while she does not believe advertisements are “immoral and unethical,” she still has “deep reservations about OpenAI’s strategy.”

“Users are interacting with an adaptive, conversational voice to which they have revealed their most private thoughts,” the researcher wrote. “People tell chatbots about their medical fears, their relationship problems, their beliefs about God and the afterlife. 

“Advertising built on that archive creates a potential for manipulating users in ways we don’t have the tools to understand, let alone prevent.”

A spokesperson for OpenAI directed The Hill to its advertising principles, which state that ads do not influence the answers ChatGPT provides, as well as CEO of Applications Fidji Simo telling journalist Alex Heath earlier this week that “ads are always going to be very clearly separate and delineated from the content, so you know exactly what you are getting.”

The Hill has also reached out to Anthropic regarding Sharma’s concerns.

Claude, Anthropic’s chatbot, released multiple ads earlier this month mocking OpenAI for its decision to display ads on ChatGPT. OpenAI CEO Sam Altman said at the Harvard Business School in 2024 that he is not “totally against” ads on chatbots, but the idea is “uniquely unsettling.”

Altman fired back at the ads — which aired during the Super Bowl — writing on social media that they are “clearly dishonest” and “deceptive.”