AI Insurance Emerges as Chatbot Failures Highlight New Liabilities

In a notable development reflecting AI’s increasing integration into business operations, insurers at Lloyd’s of London have launched specialized coverage for losses caused by artificial intelligence tool failures. This initiative, spearheaded by Armilla, a startup backed by Y Combinator, underscores growing corporate concerns about the unpredictable and costly errors AI-powered tools can generate, particularly chatbots and customer service platforms.

These innovative insurance products, underwritten by several Lloyd’s insurers, cover legal fees and damages incurred when AI tools underperform significantly, triggering customer lawsuits or third-party claims. Unlike general technology error and omission policies—which typically offer limited payouts for AI-related incidents—Armilla’s policies provide broader, AI-specific protections.

Chatbots Go Rogue

Real-world scenarios illustrate the necessity of such coverage. Earlier incidents include Virgin Money’s chatbot mistakenly reprimanding a customer for using the word “virgin,” and DPD disabling its bot after inappropriate and offensive interactions with users. Additionally, Air Canada faced financial implications when a tribunal mandated it honor a chatbot-generated false discount. According to Armilla, their policy would have absorbed such financial losses if the chatbot’s performance was demonstrably below initial benchmarks.

This targeted insurance addresses a critical gap: general technology policies typically exclude claims arising from AI’s adaptive learning processes, which inherently introduce unpredictability. Armilla’s approach assesses an AI model’s initial performance metrics and provides coverage if significant degradation occurs over time. For instance, a chatbot initially accurate in 95% of interactions dropping to 85% accuracy could trigger a claim.

Insurance Will Drive Innovation

Karthik Ramakrishnan, Armilla’s CEO, emphasizes that this insurance could encourage broader adoption of AI technologies by providing enterprises greater confidence in managing the associated risks. However, insurers, including Chaucer at Lloyd’s, clarify they will exercise selectivity, ensuring AI systems covered are initially robust and less prone to failure.

This development signals a pivotal shift for businesses and insurers alike, recognizing AI not only as a tool for innovation but also as a source of distinct operational risks requiring specialized coverage. The emergence of AI-specific insurance highlights a future where comprehensive risk management strategies must evolve alongside technological advancements, setting new standards for Integrated Risk Management practices.

[Source: Financial Times, May 11, 2025]