OpenAI warns AI is growing too fast, hires ex-Anthropic expert to manage risks

OpenAI has announced former Anthropic technical team member, Dylan Scandinaro, as its Head of Preparedness. Here is what this role means, and how much Scandinaro will earn.

The investment deepens Altman’s growing footprint across frontier technologies and sharpens his rivalry with Elon Musk.

OpenAI CEO Sam Altman announced that the AI startup had hired a new Head of Preparedness.

New Delhi,UPDATED: Feb 4, 2026 10:13 IST

OpenAI has announced that it has hired former Anthropic technical team member, Dylan Scandinaro, as its new Head of Preparedness. This new role will revolve around ensuring safeguards over the AI startup’s advancements in artificial intelligence, as it works on reaching artificial general intelligence (AGI) in the near future.

On X, OpenAI CEO Sam Altman welcomed Scandinaro’s arrival. Altman hinted that the company was making rapid advancements when it came to AI models. He added, "This will require commensurate safeguards to ensure we can continue to deliver tremendous benefits.”

Why OpenAI needed a Head of Preparedness

OpenAI first announced the vacancy for this role in December last year. At the time, Sam Altman had stated that this job would be quite stressful as the person would be tasked with leading the company’s Safety Systems team. Dylan Scandinaro will be focusing on identifying, testing, and reducing the potential harms that advanced AI models could cause.

Sam Altman added, “Dylan will lead our efforts to prepare for and mitigate these severe risks.” This hiring comes at a time when there is a lot of debate surrounding the potential risks of AI, and whether these models will remain committed to the interest of humanity in the long run.

Sam Altman X post — A screenshot of Sam Altman's post on X

The Head of Preparedness position was previously held by Aleksander Madry, who changed roles within OpenAI in July 2024.

Former OpenAI staffer Daniel Kokotajlo resigned after losing confidence in the startup’s ability to act responsibly as it moved closer to artificial general intelligence, or AGI—a still-theoretical form of AI that matches human reasoning abilities.

Who is Dylan Scandinaro, and how much will he earn at OpenAI?

Dylan Scandinaro’s latest stint was at Anthropic, where he also looked at safety over AGI development. In the past, he has also worked at companies such as Google DeepMind and Palantir.