OpenAI Advertises $555k ‘Head of Preparedness’ to Re-boot Internal Safety Unit

Big tech firms often create internal safety czars only after a jolt—the Manhattan Project’s 1943 Health Division or Facebook’s 2016 ‘Integrity’ team—when risks begin to threaten legitimacy. OpenAI’s move echoes those moments: rapid capability gains, public lawsuits over suicides, and models finding zero-days resemble early nuclear criticality accidents and 1990s Microsoft security crises, prompting hurried institutional fixes. Strategically, this hiring reveals two intersecting long-wave trends: (1) private labs are racing to codify self-regulation to pre-empt external regulation, and (2) AI companies are normalising nine-figure burn rates while selling a narrative of existential vigilance to investors and society. Whether the new Preparedness chief becomes a genuine brake on deployment or a compliance shield will shape norms for all frontier labs. On a century horizon, the episode may mark either the embryonic formation of an internal ‘IAEA for algorithms’—setting precedents for model-launch inspections—or yet another instance where commercial incentives outpace in-house overseers, as with early chemical and fossil-fuel safety offices. History suggests the durability of such safeguards will depend less on one hire and more on whether independent, enforceable governance follows.