The British government has announced the establishment of the AI Safety Institute's (AISI) first international office in San Francisco. This move aims to bolster the institute's technical capabilities and position it as a leading entity in AI safety. The new office will enable closer collaboration with key AI technology companies.
Strategic Expansion
The San Francisco office will initially recruit a Research Director to lead the first team of technical staff. This expansion will complement the London headquarters, which currently employs over 30 technical staff. Secretary of State for Science and Technology Michelle Donelan emphasized that the expansion underscores British leadership in AI safety and will allow the UK to assess AI risks and opportunities from a global perspective. She highlighted that this initiative will strengthen the UK's partnership with the US and allow other countries to benefit from British expertise in AI safety.
AISI's Recent Work
Since its inception last year, AISI has conducted a study to evaluate the effectiveness of AI safeguards. The study's findings indicate that AI safety is still in its early stages. The head of AISI noted that the results represent only a small part of the broader evaluation approach the institute is developing. The institute aims to continue its evaluations with a focus on risks linked to national security.
The expansion comes amid ongoing discussions about the importance of ensuring AI models meet safety requirements without hindering innovation. The government aims to strike a balance that allows for the development of new technologies while maintaining stringent safety standards. This approach is seen as essential, especially in light of recent controversies in the AI industry, such as those involving OpenAI.
First AI Safety Testing Results
The expansion was unveiled as the AI Safety Institute published its first-ever AI safety testing results on publicly available models and agreed on a new collaboration with Canada. The office is expected to open this summer, recruiting the first team of technical staff headed by a Research Director. By expanding its presence in the US, the Institute will establish closer collaboration with the US, furthering the country's strategic partnership and approach to AI safety, while also sharing research and conducting joint evaluations of AI models that can inform AI safety policy globally.
Also this week, AISI became controversially connected with an intelligence data firm. Faculty, a British artificial intelligence company, has secured significant contracts with the UK government, raising questions about the firm's political connections and procurement processes. The company has been awarded over £45 million in public tenders over the past three years, including a £1.5 million contract with the newly established AI Safety Institute (AISI) to test advanced AI technologies.
