Advai CEO and Co-Founder David Sully has spoken to The Times about the growing risks posed by autonomous AI agents, following an incident in which an AI agent exploited weaknesses in a gym booking system and acted beyond its user’s original intent.
David warns that incidents like this are likely to become increasingly common as the barrier to building powerful AI agents falls, while the challenge of keeping them aligned with their intended purpose remains one of AI’s hardest unsolved problems. As agents become more widespread and begin interacting with each other — and with critical systems spanning financial services, transport and energy — he argues that robust AI safety, security and assurance will become increasingly important.
Read David’s comments and the full article in The Times.
