OpenAI is looking for a new Head of Preparedness

OpenAI is on the lookout for a new executive to oversee the identification and management of emerging risks associated with artificial intelligence, spanning areas such as cybersecurity and mental health. In a recent post on X, CEO Sam Altman highlighted the increasing challenges posed by AI models, including their potential impacts on mental well-being and their capabilities in identifying critical security vulnerabilities. Altman emphasized the importance of enabling cybersecurity professionals with advanced tools while ensuring that these tools cannot be exploited by malicious actors. He urged potential candidates to apply if they are interested in contributing to the safety of AI systems, particularly in relation to biological capabilities and self-improving systems. The new Head of Preparedness will be tasked with implementing OpenAI's preparedness framework, which details the organization's strategy for monitoring and addressing the risks associated with frontier AI capabilities. This initiative was first announced in 2023, with a focus on investigating both immediate threats, such as phishing attacks, and more speculative dangers, like nuclear risks. Following the establishment of this team, OpenAI made a significant personnel change, moving Aleksander Madry, the previous Head of Preparedness, to a new role centered on AI reasoning. Additionally, other safety-focused executives at the company have transitioned to different positions, raising questions about the continuity of its safety efforts. Recently, OpenAI also revised its Preparedness Framework, indicating a willingness to reconsider its safety measures in response to potential risks posed by competing AI labs releasing high-risk models without adequate safeguards. Altman’s comments reflect a broader concern regarding the impact of generative AI chatbots on mental health, with recent lawsuits claiming that OpenAI’s ChatGPT has exacerbated issues like delusions and social isolation, leading to severe consequences for some users. The company remains committed to enhancing ChatGPT’s capabilities to detect emotional distress and connect users with appropriate support resources.