In a significant development for the rapidly evolving AI industry, Lloyd's of London has launched a pioneering insurance product through startup Armilla that specifically addresses the financial risks associated with artificial intelligence malfunctions.
The new policy, officially launched on April 30, 2025, is underwritten by several Lloyd's syndicates including Chaucer Group. It provides explicit coverage for businesses facing legal claims when AI tools generate inaccurate outputs or underperform, covering costs such as court-awarded damages and legal expenses.
Karthik Ramakrishnan, CEO of Armilla, explains the product's significance: "Businesses are racing to deploy AI, but their risk management and insurance tools haven't kept pace. There's a growing concern of 'silent AI cover' – the uncertainty of whether existing policies will respond to AI-specific failures."
The insurance specifically targets risks like AI hallucinations (when models confidently present fabricated information), deteriorating model performance, and deviations from expected behavior. Unlike traditional technology policies that often impose low caps on AI-related liabilities, Armilla's coverage activates only when an AI system's performance falls significantly below pre-established benchmarks.
This development comes amid increasing corporate adoption of AI technologies and several high-profile incidents where AI chatbots have generated misleading or harmful outputs. In one notable case, Air Canada was required to honor discounts mistakenly offered by its chatbot – a loss that Armilla claims would have been covered under its policy had the chatbot been deemed to perform below expectations.
Industry experts view this insurance product as a significant milestone in the maturing of the AI sector. With the global AI insurance market projected to grow at a compound annual growth rate of 33.06% between 2024 and 2034, Lloyd's and Armilla's collaboration represents an evolutionary step in insurance to accommodate the demands of a digitizing economy while promoting responsible AI adoption.
As Tom Graham of Chaucer notes, insurers will be selective: "We won't cover AI systems that are excessively prone to breakdown." This approach incentivizes companies to maintain high AI standards while providing the financial safeguards necessary for broader AI deployment.