Blighty and the United States have sealed a historic pact to pioneer AI safety testing, becoming the globe’s inaugural nations to formalise a collaborative agreement in assessing the threats of artificial intelligence.
The accord, signed by Britain’s tech secretary Michelle Donelan and US commerce supremo Gina Raimondo, outlines a course for joint efforts and pooling expertise, data, and insight.
The duo intends to construct a “unified strategy” for AI safety testing, conducting “at least” one shared testing drill on a model accessible to the public.
Under the freshly minted AI Safety Institute in the UK and its counterpart, yet to be operational in the US, the agencies will facilitate researcher exchanges and craft methodologies to autonomously scrutinise private AI systems erected by tech titans such as Google, Oracle, and OpenAI.
Drawing inspiration from the historical liaison between GCHQ and the National Security Agency predating the Second World War, this partnership mirrors that esteemed alliance.
“We’ve been unequivocal that ensuring AI’s safe progression is a global responsibility,” Donelan affirmed. “Only through concerted effort can we confront the per cent perils of this technology and unlock its boundless potential to enrich our lives.”
“Our collaborative endeavours in AI safety will fortify the groundwork we laid at Bletchley Park last November, and I’m confident our combined prowess will continue to lead nations toward responsibly harnessing AI’s vast benefits,” she added.
Donelan reiterated the UK’s stance against AI regulation, citing the breakneck pace of technological advancement.
This stark contrast with other jurisdictions, including the US, looms large. President Biden issued a targeted executive order last year, while the EU’s AI Act stands as the world’s sternest regulation governing this terrain.
While this landmark accord marks a premiere, both nations have also pledged to forge similar alliances with other nations to propagate AI safety worldwide.
