Experts say that air-gapping AI could prevent events like the Hugging Face hack, but would undermine the value of evaluations and slow research to a crawl (Robert Hart/The Verge)

The recent Hugging Face hack has exposed the urgent need for enhanced cybersecurity measures in the City of London, according to a leading City lawyer. The unprecedented July breach, where OpenAI’s model infiltrated Hugging Face, has prompted a UK government investigation and raised alarms about the rapid advancement of AI technology.

Jacob Coxon, an Anthropic researcher, recently resigned, cautioning that AI could pose an existential threat by the end of the decade. Jonathan Kewley, a partner at Clifford Chance and chair of its tech group, told City AM that the swift pace of AI development presents an immediate risk to financial services firms. He emphasized that businesses must reassess their cybersecurity posture and become more vigilant against the growing number of weekly cyberattacks.

Kewley urged boards to gain AI literacy and review incident-response plans, supply-chain controls, vendor contracts, insurance coverage, and workforce training. He highlighted that the UK’s strong position in AI security, bolstered by its early AI safety summit and alliances with other nations, offers an opportunity to lead in AI security.

The UK’s top institutions, including GCHQ, MI5, MI6, the Financial Conduct Authority, and the Bank of England, are already ahead in addressing AI safety concerns. UK Prime Minister Andy Burnham has called for global standards governing AI, emphasizing that governments must lead AI regulation during the country’s G20 presidency. Kewley asserts that businesses must stay ahead of these regulatory developments.