OpenAI’s AGI Leader Departs, Raising Concerns About Readiness for Human-Level AI

In a significant exit, Miles Brundage, OpenAI’s senior adviser for AGI (artificial general intelligence) readiness, announced his departure, stating neither OpenAI nor any other major AI lab is adequately prepared for AGI. Brundage, who played a key role in shaping OpenAI’s safety efforts, believes that more independent voices are essential for unbiased AI policy and governance. His exit underscores a shift in OpenAI’s focus from long-term safety to product-driven goals, a shift that also recently led to the dissolution of OpenAI’s Superalignment and AGI Readiness teams.

Diverging Paths: Commercial Ambitions and AGI Safety

Tensions within OpenAI’s mission emerged in recent years, particularly as it moved toward for-profit status. The pressures from a recent $6.6 billion investment round underscore these changes, aligning OpenAI more with product innovation than with the safety research many of its researchers originally joined to advance. Brundage's exit is the latest in a string of high-profile departures from OpenAI’s safety divisions, including Jan Leike and cofounder Ilya Sutskever, both of whom expressed concerns about OpenAI’s shift away from its original mission.

A Call for External Oversight in AI

Leaving OpenAI, Brundage aims to contribute to global AI governance outside the tech industry, free from commercial constraints. OpenAI has expressed support for his independent work, offering funding and resources with no attached conditions.

As OpenAI approaches critical AGI milestones, these events illustrate a challenging balance between rapid AI development and ensuring the safety and alignment of powerful AI models, suggesting a pressing need for industry-wide collaboration and regulatory oversight.