Archive · 10 Sep 2026
AI Safety Researchers Leave Labs Over Humanity Risks
Several AI researchers have left major technology companies because they are worried about the speed of AI development.
They say companies may be focusing too much on releasing products and competing with one another.
Jacob Coxon warned that future AI systems could become extremely powerful and difficult to control.
He said some researchers believe these systems could seriously harm humanity.
Other former OpenAI and Anthropic employees have made similar criticisms.
They have called for stronger safety testing, more oversight, and slower development.
Some researchers left voluntarily, while others were fired after alleged information leaks.
The companies’ race to build more advanced AI is at the center of the debate.
Jacob Coxon resigned from Anthropic and warned that companies are irresponsibly racing toward self-improving superintelligence.
Coxon said some AI researchers privately fear advanced systems could pose an existential danger by the end of the decade.
Former Anthropic safeguards head Mrinank Sharma cited friction between safety checks and aggressive commercial deployment schedules.
OpenAI safety leaders including Ilya Sutskever, Jan Leike, Daniel Kokotajlo, and William Saunders have also departed after raising concerns.
Other departures included policy lead Gretchen Krueger, while Leopold Aschenbrenner and Pavel Izmailov were fired over alleged confidential-information leaks.
- Who
- AI researchers and policy leaders, including Jacob Coxon, Ilya Sutskever, Jan Leike, Daniel Kokotajlo, William Saunders, Mrinank Sharma, and Gretchen Krueger.
- What
- Researchers have resigned or been dismissed from OpenAI and Anthropic while warning that commercial pressure and rapid AI development may create serious risks.
- When
- The departures discussed occurred mainly over the past two years, including resignations in 2024 and February 2026.
- Where
- At major AI companies, particularly OpenAI and Anthropic.
- Why
- The researchers cited concerns about insufficient safety work, weak oversight, commercial pressure, and an uncontrolled race toward advanced AI systems.
- Latest resignation
- Jacob Coxon resigned from Anthropic and publicly criticized the AI industry’s approach to safety.
- Coxon’s warning
- He said AI companies are racing toward self-improving superintelligence and called the situation a gamble with human lives.
- Anthropic departure
- Mrinank Sharma, the former head of Anthropic’s safeguards research team, resigned in February 2026.
- OpenAI departures
- Ilya Sutskever, Jan Leike, Daniel Kokotajlo, William Saunders, and Gretchen Krueger are among those who left OpenAI after raising safety or governance concerns.
- Dismissed researchers
- Leopold Aschenbrenner and Pavel Izmailov were fired by OpenAI for allegedly leaking confidential information.
- Central concern
- Former researchers said safety checks and oversight were being overshadowed by commercial deployment and product development.