Feds appoint “AI doomer” to run US AI safety institute

The US AI Safety Institute, a part of the National Institute of Standards and Technology (NIST), has named its leadership team, with Paul Christiano appointed as the head of AI safety. Christiano, known for predicting potential AI-related risks, has faced criticism for his views that there is a significant chance of AI development leading to disastrous outcomes. Despite concerns among some NIST staffers regarding Christiano’s appointment, the Commerce Department has emphasized the importance of bringing in top talent to address AI safety issues in line with the institute’s mission to enhance economic security and quality of life.

Christiano’s background includes creating foundational AI safety techniques and founding the Alignment Research Center (ARC) to align machine learning systems with human interests. His role involves overseeing evaluations of frontier AI models and implementing risk mitigations to enhance model safety and security. While some critics have expressed reservations about Christiano’s outlook on AI risks potentially overshadowing present ethical concerns regarding AI, supporters believe his expertise makes him well-suited for the position.

In an effort to address potential risks associated with rapidly advancing AI technologies, Christiano has emphasized the importance of responsible scaling policies and regulations to manage the increasing complexity of AI systems. Despite calls from some quarters for a temporary pause in AI development to improve protective measures, Christiano argues that the current level of risk is manageable with appropriate detection and response plans in place.

The leadership team of the AI Safety Institute also includes individuals with diverse expertise in areas such as human-AI teaming, international engagement, and ethical AI research. The institute aims to safeguard against AI-related risks while harnessing the benefits of AI technology, underlining the critical role of effective leadership in navigating the evolving landscape of AI safety and security.