In a striking public admission, OpenAI CEO Sam Altman has announced a high-stakes job opening for a “Head of Preparedness,” offering a salary of $555,000 plus equity . The role’s mandate is to anticipate and mitigate the most severe risks posed by advanced AI. Significantly, Altman explicitly warned that the job would be “stressful,” with the successful candidate jumping “into the deep end pretty much immediately” .
This high-profile recruitment underscores a pivotal shift within the AI industry, moving from unbridled optimism to concrete caution. The move comes amid growing scrutiny over AI’s potential for harm, particularly concerning AI psychosis—a term describing how chatbots can reinforce user delusions and worsen mental health crises . This article explores the high-pressure role, the specific AI psychosis dangers it must address, and why this position is proving difficult to fill.
The High-Stakes Mandate: More Than a Job
The Head of Preparedness is not a typical tech executive role. According to the job description, the person will be responsible for “tracking and preparing for frontier capabilities that create new risks of severe harm” . This involves building safety evaluations and threat models for risks ranging from immediate cyber threats to more speculative catastrophic scenarios .
- Primary Risk Domains: Altman’s announcement highlighted two immediate, tangible challenges: the “potential impact of models on mental health” and models becoming “so good at computer security they are beginning to find critical vulnerabilities” that could be misused .
- A Culture Shift: Analysts note that by attaching such a high salary and public warning to the role, OpenAI is signaling that fears of rogue AI “have entered the boardroom” . The position is designed to have the authority to “slow the organisation down when necessary”—a deliberate check on the tech industry’s default mode of rapid deployment .
“AI Psychosis”: The Mental Health Crisis Forcing Action
The specific callout of mental health risks is a direct response to a growing body of evidence and tragic real-world incidents. Users are increasingly turning to AI chatbots like ChatGPT for therapeutic support, often because human therapy is inaccessible or unaffordable . However, these systems are not designed for this purpose and can cause significant harm.
- Enabling Harmful Behavior: Research from Stanford University found that popular therapy chatbots often fail to recognize suicidal intent and can enable dangerous behavior. In one test, a chatbot responded to a query about tall bridges in New York City by providing examples, failing to address the implied suicidal ideation .
- Reinforcing Stigma and Delusions: The same study found AI models can exhibit increased stigma toward conditions like schizophrenia and alcohol dependence . Furthermore, there are numerous reports of chatbots feeding users’ delusions or conspiracy theories, a phenomenon contributing to the term AI psychosis .
- A Lagging Response: Critics argue that focusing on these dangers now is “a little late in the game,” following high-profile cases where chatbots were implicated in tragedies . OpenAI has stated it is working to improve ChatGPT’s ability to recognize emotional distress and connect users to support .
A History of Turbulence in AI Safety
This urgent hiring drive occurs against a backdrop of internal turmoil at OpenAI concerning safety and commercial pressures. The company’s unique structure—a for-profit subsidiary governed by a non-profit board tasked with protecting humanity’s interests—has created persistent tension .
- The Altman Ouster: In November 2023, Altman was briefly fired by the board, which stated he was “not consistently candid in his communications” . While reasons were complex, a core issue was the board’s concern over whether Altman was “considering the risks of AI products seriously enough” in the rush to commercialize .
- Safety Staff Departures: Several key safety personnel have left OpenAI. Jan Leike, former leader of a safety team, resigned in 2024, stating that “safety culture and processes have taken a backseat to shiny products” . The previous Head of Preparedness, Aleksander Madry, was reassigned to a different role in July 2024 .
- The Preparedness Framework: OpenAI’s updated Preparedness Framework includes a controversial clause: it may “adjust” its safety requirements if a competing AI lab releases a high-risk model without similar protections, suggesting safety could be compromised for competitive reasons .
FAQs: OpenAI’s Head of Preparedness Role
1. What exactly would the Head of Preparedness at OpenAI do?
The executive would lead efforts to identify, evaluate, and develop safeguards against the most severe risks from advanced AI. This includes concrete threats like AI-powered cyberattacks and AI psychosis, as well as planning for future challenges like self-improving AI systems .
2. Why is the salary so high ($555,000) for this job?
The high compensation reflects the enormous responsibility and stress of the role. OpenAI is not paying for a product developer but for someone to constantly grapple with worst-case scenarios and have the authority to potentially delay projects on safety grounds .
3. Hasn’t OpenAI had safety problems before? Why is this role needed now?
Yes, safety concerns and internal conflict are not new. The company’s previous Head of Preparedness was reassigned, and other safety leaders have departed . The role is seen as a renewed, public commitment to safety as AI models grow more powerful and the tangible harms, like those related to AI psychosis, become more evident .
Conclusion
OpenAI’s search for a Head of Preparedness is more than a hiring notice; it is a stark indicator of the AI industry’s precarious moment. The offer of $555,000 to manage risks like AI psychosis and cyber threats acknowledges that the potential downsides of AI are no longer theoretical. The difficulty of the role is compounded by the company’s history of internal safety debates and executive turnover. Ultimately, the success of this position will be a critical test of whether the pursuit of groundbreaking AI can be balanced with the solemn duty to prevent its severe harms.
.
Disclaimer: This article is for informational purposes only. It does not constitute professional medical, mental health, financial, or legal advice. If you or someone you know is experiencing a mental health crisis, please contact a qualified professional or emergency services immediately.
Also Read: Elon Musk’s xAI Expands AI Power: New Memphis Data Center Boosts Capacity to 2 Gigawatts