Having previously discussed possible ideas for global AI regulation with OpenAI chief executive Sam Altman last week, Mr Sunak aims to secure involvement of British “leadership” in a prospective framework that other nations and regions could follow.
As legislation is being drafted in regions such as China, the European Union and the state of California, pending discussions with President Biden are set to cover ways to implement guardrails through collaboration with global allies.
It’s reported that the UK government plans for AI research and regulation are inspired by bodies including the International Atomic Energy Authority (IAEA) and the European Organisation for Nuclear Research (CERN).
AI development needs state control, says pioneer — Yoshua Bengio, one of the three ‘godfathers of AI’, says firms developing artificial intelligence systems need government oversight.
A ‘CERN for AI’ had previously been mooted globally, and an unnamed ally of Mr Sunak told the FT that “agreeing a joint approach and ensuring international alignment” was the first port of call.
According to former UK Chancellor Sajid Javid, who has previously served as an adviser for US-based firm C3 AI, both the CERN and IAEA models were complementary to Sunak’s plans, with CERN particularly believed to provide a “very controlled environment” for global development.
As for the Vienna-based IAEA, its standing as an international regulator and surveillance body for responsible nuclear energy usage has been discussed, with Javid stating that “you need a regulator and also ethical research backed by western governments”.
OpenAI CEO Altman, while looking to secure operations in Europe in the near future, has shared concerns around the regulation — he commented that if this was achieved, OpenAI would leave Europe in the case that its technology was deemed ‘high-risk’ by the EU by 2025, before backtracking and stating the start-up “had no plans to leave”.
San Francisco-based Anthropic — the developer of transformer-based large language model Claude — meanwhile, has opted to establish a UK office in the aim of rivalling OpenAI’s ChatGPT.
Related:
EU/US global AI code of conduct ‘within weeks’ — European Union and United States working at speed to publish voluntary code of conduct for companies working in artificial intelligence, ahead of legislation.
