BLETCHLEY PARK, England (AP) — Digital officials, tech company bosses and researchers are converging Wednesday at a former codebreaking spy base near London to discuss and better understand the extreme risks posed by cutting-edge artificial intelligence.
The two-day summit focusing on so-called frontier AI notched up an early achievement with officials from 28 nations and the European Union signing an agreement on safe and responsible development of the technology.
Frontier AI is shorthand for the latest and most powerful general purpose systems that take the technology right up to its limits, but could come with as-yet-unknown dangers. They’re underpinned by foundation models, which power chatbots like OpenAI’s ChatGPT and Google’s Bard and are trained on vast pools of information scraped from the internet.
The AI Safety Summit is a labor of love for British Prime Minister Rishi Sunak, a tech-loving former banker who wants the U.K. to be a hub for computing innovation and has framed the summit as the start of a global conversation about the safe development of AI. But U.S. Vice President Kamala Harris may divert attention Wednesday with a separate speech in London setting out the Biden administration’s more hands-on approach.