Janice R. Andrews
A senior safety researcher at OpenAI, Steven Adler, has stepped down from his position, expressing grave concerns about the rapid advancement toward artificial general intelligence (AGI). Adler shared his worries on social media platform X, characterizing the ongoing competition between AI laboratories as highly dangerous and potentially catastrophic for humanity.
In his public statement, Adler emphasized that no AI lab has currently solved the challenge of AI alignment – ensuring AI systems operate in harmony with human values. He stressed that the accelerating pace of development could make finding such solutions increasingly difficult.
Drawing from his four-year tenure at OpenAI, Adler shared personal reflections about humanity’s future prospects, questioning long-term life decisions in light of AI development risks. His position aligns with other prominent voices in the field, including UC Berkeley’s Professor Stuart Russell, who compared the AGI race to heading toward a cliff’s edge.
This departure adds to a series of exits from OpenAI’s safety team, including notable figures Ilya Sutskever and Jan Leike. Leike criticized the company’s focus on product development over safety considerations. According to former governance researcher Daniel Kokotajlo, approximately half of OpenAI’s long-term safety team members have departed.
The situation intensified following reports that Chinese AI firm DeepSeek developed a competitive model at lower costs. In response, OpenAI’s CEO Sam Altman announced faster product releases, describing the competition as “invigorating.”
Industry experts are calling for stronger regulatory oversight to prevent dangerous acceleration in AI development. While Adler advocates for enhanced safety protocols, questions remain about whether companies will moderate their pace given the intense global competition.
