Why Insurers Are Pulling Back from AI Risk
Artificial intelligence (AI) is revolutionising industries, but insurers are warning that AI may now be too unpredictable to insure. This could leave UK businesses exposed if widely used models fail at scale; an AI risk insurance UK
What’s Driving Insurer Anxiety?
Recent reports in the Financial Times reveal growing concern as companies deploy generative AI for customer service, product design, and cybersecurity. Major US insurers – including Great American, Chubb, and W. R. Berkley – have asked regulators to exclude AI-related liabilities from standard policies. The fear? Systemic risk – one failure triggering thousands of claims.
The Core Challenge: Unpredictable AI Behaviour
Actuaries rely on historical data to model risk. Generative AI, however, is a “black box” with limited transparency. Insurers cannot predict how often errors will occur or how severe they might be. A single flawed update could create thousands of identical losses – something the insurance market cannot absorb.
Real-World AI Failures Highlight the Risk
- Google AI Overview falsely accused a solar company of legal violations, sparking a £90m lawsuit.
- Air Canada was held liable after its chatbot invented a discount policy.
- Arup lost £25m after fraudsters used a deepfake to authorise a transfer.
Cybersecurity reports show AI-assisted phishing and deepfake fraud are rising sharply. The UK’s National Cyber Security Centre warns that AI lowers the barrier for criminals to launch convincing scams.
Why Insurers Are Seeking AI Exclusions
Filings in the US show insurers requesting permission to exclude claims linked to “any actual or alleged use” of AI. Mosaic Insurance, for example, covers certain embedded AI software but excludes general-purpose models like ChatGPT.
Regulatory and Industry Response
- The Geneva Association warns parts of AI risk may become “uninsurable” without better transparency.
- Lloyd’s of London has urged underwriters to review AI exposure in cyber policies.
- EU AI Act introduces strict rules for high-risk systems, requiring risk assessments and transparency measures.
- UK regulators have taken a lighter, sector-based approach, leaving more responsibility with businesses.
What This Means for UK Businesses
If insurers exclude AI-related losses, businesses face gaps in cover. Many UK firms already use AI for customer service, content creation, coding, and HR screening. A chatbot error or flawed design could leave a company exposed without liability protection.
Action Points for UK Organisations
- Audit your AI usage – Identify where liability could arise.
- Strengthen governance – Maintain documentation and audit trails.
- Monitor regulations – EU AI Act may influence UK standards.
- Plan for uninsured risk – Consider contractual safeguards and internal controls.
Future Outlook: Specialist AI Insurance?
Analysts predict new AI insurance products will emerge, similar to cyber cover. But meaningful protection may remain elusive until insurers gain visibility into how models are trained and behave under stress.
Ready to Protect Your Business?
Contact EC Computers today for a cyber security assessment and discover how we can help you stay secure.
📞 Call us: 0117 200 1000
📧 Email: Contact-us form
🌐 Visit: https://eccomputers.co.uk/cyber-security/
Further reading: Managed IT Services and Support – Keeper Password Manager
Importance of strong passwords – Backup of Microsoft 365 external to inbuilt recovery – Privacy compliance and new 2025 laws
#ArtificialIntelligence #GenerativeAI #MachineLearning #DigitalTransformation #TechInnovation #FutureOfWork #AIRegulation #CyberSecurity #DataPrivacy #FinTech