Jan Leike, previously the head of alignment at OpenAI and leader of the ‘Superalignment’ team, has stepped down from his position. Leike announced his resignation on May 17 via a series of posts on the social media platform X, formerly known as Twitter. He expressed concerns that OpenAI was prioritizing product development over crucial AI safety measures.

Leike, who had been with OpenAI for about three years, criticized the company for neglecting AI safety in favor of developing attractive products. He pointed out that his team lacked the necessary resources, particularly computing power, to conduct essential safety research. Leike emphasized that creating machines smarter than humans is inherently risky, underscoring OpenAI's responsibility to prioritize safety.

Leike's departure coincided with the resignation of Ilya Sutskever, OpenAI’s chief scientist and co-leader of the ‘Superalignment’ team. Sutskever, a co-founder of OpenAI and a key figure in developing ChatGPT, had announced his resignation just days earlier. Their departures have led to significant changes within OpenAI.

In response to these resignations, OpenAI decided to disband the ‘Superalignment’ team, integrating its functions into other research projects. This move comes amid ongoing internal restructuring following the governance crisis in November 2023, when CEO Sam Altman was temporarily ousted and President Greg Brockman lost his chairmanship.

The ‘Superalignment’ team was formed to address the existential risks posed by advanced AI systems and develop methods to control and steer superintelligent AI. Despite the team's dissolution, OpenAI has pledged to continue long-term AI risk research under John Schulman, who also leads a team focused on fine-tuning AI models post-training.

The resignations and restructuring have heightened scrutiny of AI safety and governance at OpenAI. This follows a period of internal disagreement, particularly after Altman's brief dismissal and subsequent reinstatement. The developments suggest that OpenAI may be prioritizing rapid AI advancement over safety.

Recently, OpenAI launched a new multimodal AI model, GPT-4o, capable of more natural and human-like interactions. While this achievement showcases OpenAI's technological prowess, it also raises ethical concerns about privacy, emotional manipulation, and cybersecurity.

Despite the upheaval, OpenAI maintains its goal of developing AGI safely and for the benefit of humanity. In a post on X, CEO Sam Altman acknowledged Leike's contributions and reaffirmed the company's commitment to AI safety. "I’m very grateful to @janleike for his great contributions to OpenAI’s alignment research and safety culture, and I am really sad that he is leaving. He’s right we have a lot more work to do; we are determined to do it. I will post my longer version in the next couple of days,” Altman wrote.

Leike and Sutskever's exits, along with the dissolution of the ‘Superalignment’ team, highlight ongoing tensions within OpenAI regarding its priorities. As the company continues to develop advanced AI models, the debate over balancing innovation with safety remains critical.