OpenAI Hires Anthropic’s Dylan Scandinaro as Head of AI Preparedness

OpenAI Bolsters AI Safety Team with Anthropic Hire

OpenAI has appointed Dylan Scandinaro, formerly of Anthropic, as its new Head of Preparedness. The move signals a heightened focus on safety oversight as the company accelerates development in artificial intelligence, including progress toward artificial general intelligence (AGI).

The appointment was announced by OpenAI CEO Sam Altman on X, who indicated the company’s rapid advancements necessitate “commensurate safeguards to ensure we can continue to deliver tremendous benefits.”

A Role Born of Growing AI Concerns

OpenAI initially advertised the Head of Preparedness position in December 2024, with CEO Sam Altman describing it as a “high-stress” role. The individual selected would be responsible for leading the company’s Safety Systems team, focusing on identifying, testing, and mitigating potential harms associated with increasingly advanced AI models.

Altman emphasized the need to “prepare for and mitigate these severe risks,” a sentiment reflecting ongoing debate surrounding the potential long-term implications of AI and its alignment with human interests.

The position was previously held by Aleksander Madry, who transitioned to a different role within OpenAI in July 2024.

The creation of this role and subsequent hiring comes after former OpenAI staffer Daniel Kokotajlo resigned, reportedly due to a loss of confidence in the company’s ability to act responsibly as it approaches AGI—a theoretical level of AI capable of human-level reasoning.

Who is Dylan Scandinaro?

Scandinaro joins OpenAI from Anthropic, where he also concentrated on safety related to AGI development. His prior experience includes roles at Google DeepMind and Palantir.

The Head of Preparedness position carries a substantial compensation package, with a listed annual salary of at least $555,000 (approximately Rs 4.16 crore).

Scandinaro announced his move to OpenAI on X.