OpenAI is looking for a new head of AI risk preparedness
OpenAI has opened a vacancy for the position of head of the risk assessment department. The position involves analyzing and identifying threats related to the capabilities of advanced AI models, particularly in the areas of cybersecurity and mental health.
What is known
In a post on X, OpenAI CEO Sam Altman wrote that current AI models are 'starting to pose serious challenges.' He mentioned the capability of models to detect critical vulnerabilities in security systems, as well as their potential impact on users' mental states.
The job description states that the head will implement the Preparedness Framework – the company's internal strategy for tracking and preparing for risks that may pose a threat. The salary is $555,000 per year plus options.
The Preparedness team was created in 2023 to explore potentially dangerous scenarios – from phishing attacks to threats related to biotechnology. In less than a year, the first head of this direction, Alexander Madry, moved to the team working on AI reasoning. Other heads of security departments also changed directions or left the company.
OpenAI also updated its approach to security, noting that it might change its requirements if other labs release 'high-risk' models without appropriate protective measures.
Altman's post also mentioned concerns about some AI models' impact on mental health. Previous lawsuits claimed that ChatGPT exacerbated delusions, contributed to social isolation, and was involved in tragic incidents. The company reports working on improving the recognition of emotional states and redirecting users to support services.
Source: Sam Altman
