The UK government’s AI Safety Institute (AISI) is to open a new office in San Francisco to enable it to tap global talent on both sides of the Atlantic. This news comes as the second international AI Safety Summit is to be held this week in Seoul, South Korea, where governments are expected to press tech firms on AI safety.

AISI was set up a year ago as a research organisation within the UK government’s Department of Science, Innovation and Technology. Its mission is to measure AI safety risks so as to equip governments with an empirical understanding of the safety of advanced AI systems.

Ian Hogarth, chair of the AI Safety Institute, said in a post on AISI’s website that there was “a strong belief that AI is a transformational technology that has an incredible capacity to improve our lives... With such a fast-moving technology premature or excessively broad regulation can damage innovation.”

He points out, however, that “optimism does not mean blindly ignoring risks”. AISI is assessing...