OpenAI's $555K Head of Preparedness Role Faces Hiring Challenges

OpenAI is currently seeking a new head of preparedness, a role that offers a base salary of $555,000 plus equity. While the compensation is competitive, industry experts suggest that filling the position may prove difficult due to the inherent conflicts between safety protocols and rapid product development.

The individual hired for this role will be responsible for balancing AI safety risks against the fast-paced release schedule of CEO Sam Altman. Altman has personally described the position as a 'stressful' job where the new hire will 'jump into the deep end pretty much immediately.' The challenge lies in the necessity to potentially slow down product rollouts to ensure safety, a task one expert described as 'rolling a rock up a steep hill.'

Recruiting for the head of preparedness role is proving to be trickier than it might seem, despite the significant financial package. The core of the difficulty lies in the job's requirement to navigate the tension between ensuring AI safety and supporting the company's aggressive growth strategy. Sam Altman has demonstrated a strong preference for releasing products at an exceptionally fast pace.

This year alone, OpenAI has launched several major products, including:

• Sora 2 video app
• Instant Checkout for ChatGPT
• New AI models and developer tools
• More advanced agent capabilities

According to Maura Grossman, a research professor at the University of Waterloo's School of Computer Science, the role is 'close to an impossible job.' She noted that the person in this position will likely need to tell Altman to slow down or that specific goals should not be met. Grossman characterized this effort as 'rolling a rock up a steep hill.' Despite these hurdles, the role remains a potential 'dream come true' for the right individual, given OpenAI's massive impact on daily life and the competitive compensation standards for AI talent.

The job posting for the head of preparedness does not list traditional requirements such as a college degree or a minimum number of years of experience. Instead, OpenAI specifies that a candidate 'might thrive' if they possess specific attributes. These include experience leading technical teams and the ability to make high-stakes technical judgments under uncertainty.

Key qualifications sought include:

• Deep technical expertise in machine learning, AI safety, evaluation, or security
• Ability to align diverse stakeholders around safety decisions
• Comfort with high-stakes judgment under uncertainty

There is differing opinions on the ideal background for the hire. Richard Lachman, a professor of digital media at Toronto Metropolitan University, suggests that a seasoned tech-industry executive might be a better fit than an academic. He argues that academic types tend to be more cautious and risk-averse. Lachman expects OpenAI to seek someone who can protect the company's public image regarding safety while allowing for continued rapid innovation. He described the ideal candidate not as a 'yes person,' but as someone who is 'on brand.'

This recruitment drive follows the departure of Aleksander Madry, the former head of preparedness, who moved into a new role in July 2024. Madry left a vacancy within the company's Safety Systems team, which is responsible for building evaluations, safety frameworks, and safeguards for AI models. His background was in academia, which has sparked the debate regarding the ideal profile for his successor.

OpenAI's approach to safety has been a point of contention. The company has faced internal concerns that prompted resignations from prominent early employees, including a former head of its safety team. Externally, the company has been sued by individuals alleging that its technology reinforces delusions and drives harmful behavior.

In October, OpenAI acknowledged that some ChatGPT users have exhibited possible signs of mental health problems. The company stated it is working with mental health experts to improve the chatbot's responses to users showing signs of psychosis, mania, self-harm, suicide, or emotional attachment.