One less in the safety faction: OpenAI's Chief Futurist resigns, and the alignment team has also been disbanded.

OpenAI’s Chief Futurist Joshua Achiam has announced that he is stepping down. The “mission alignment” team he once led was disbanded earlier this year in February, and former White House adviser Dean Ball takes over this week.
(Background: Google’s security director speaks with anguish: management has “lost moral principles” by abandoning its carbon-neutrality commitments and signing a confidential AI contract with the U.S. Department of War.)
(Background information: AI Renaissance — philosophers become sought-after commodities in AI labs, with ethics written into your models.)

Table of Contents

Toggle

• From “Mission Alignment” to “Chief Futurist”
• The List of Departing Safety Faction Members Grows Longer
• The Distance Between the Giant and the Mission

For nearly nine years, Joshua Achiam was at OpenAI—rising from intern all the way to Chief Futurist. But this Tuesday, he announced that he would leave later this month. The timing is sensitive: the company is preparing closely for an IPO, and he is yet another name disappearing from the list of top executives in the safety faction.

In an internal letter to staff, he wrote: “The world already knows this secret. It feels possible to keep pushing this mission from beyond the high walls of the frontier lab.”

From “Mission Alignment” to “Chief Futurist”

Achiam joined OpenAI in 2017 as an intern, and later became a research scientist focused on AI safety. Internally, he was regarded as a steadfast defender of the safety mission, but he was also controversial for occasionally criticizing the broader AI safety community.

In 2024, OpenAI set up the “Mission Alignment Team,” led by Achiam. Its mission was to ensure the company’s words and actions stayed consistent. This February, the team was disbanded, and Achiam moved into a newly created role of Chief Futurist. The position sits at the intersection of the AI safety and policy teams, researching the potential harms and benefits brought by the rise of AI. He worked with senior executives such as Chris Lehane, Chief of Global Affairs, advocating for government regulation aligned with OpenAI’s mission.

His successor, former White House AI adviser Dean Ball, took office this week. His title is “Director of Strategic Foresight,” and he will briefly overlap with Achiam during the handover. One team disbands and is restructured, one person comes and another goes—OpenAI, it seems, has not yet found a stable answer to the question of “who will guard the mission.”

The List of Departing Safety Faction Members Grows Longer

Achiam is neither the first nor the last. In 2024, Jan Leike, who co-led the “mission alignment” team with him, left to join Anthropic. Before he departed, he publicly criticized the company’s “safety culture and processes” for having made way for “eye-catching products.”

In the same year, Policy Research Director Miles Brundage and Steven Adler—who researched dangerous capabilities of AI models—also left in succession. Each went on to establish a nonprofit organization advocating that AI labs comply with safety standards. By the end of 2025, Andrea Vallone, who researched how ChatGPT responds to users troubled by psychological and emotional difficulties, also left to join Leike’s team at Anthropic.

Put these names together and the direction is clear: OpenAI’s safety talent are not being laid off—they are choosing to leave on their own, and most are heading to the same place: Anthropic, or self-founded nonprofit organizations. In short, the willingness of these people to keep pushing for safety from within OpenAI is declining.

The Distance Between the Giant and the Mission

OpenAI is no longer the small research lab it was in 2017. Since the launch of ChatGPT in 2022, the company’s scale and commercial pressure have grown in step, with safety, product, and research teams being repeatedly reorganized. And now, it is preparing for an IPO, moving toward becoming a true technology giant. In his letter, Achiam said he believes “we can reach a world of peace, unprecedented prosperity, and possibilities beyond imagination,” and he pledged that no matter what he does next, he will continue working toward that vision.

But the question is whether this vision requires someone to push for it from inside the walls, or whether it can be realized from outside. Achiam chose the latter. His answer, to some extent, is also the collective answer the entire safety faction has been giving over the past few years: when the engine of commercialization runs at full speed, staying inside the high walls to guard the mission no longer seems to them to be the most effective position.