AI Is Too Risky to Insure, Say People Whose Job Is Insuring Risk

Why It Matters

The move signals a looming gap in risk management for AI deployments, forcing companies to shoulder potential liabilities and potentially slowing AI adoption across industries.

Key Takeaways

• Insurers request AI liability exclusions from U.S. regulators
• Recent AI errors cost $110M and $25M lawsuits
• Systemic AI failures could trigger thousands of simultaneous claims
• Current policies cover single $400M loss, not mass incidents
• Coverage gap may slow corporate AI deployment

Pulse Analysis

The insurance industry is confronting a paradigm shift as artificial intelligence moves from experimental labs to core business functions. Traditional underwriting models rely on historical loss data and predictable exposure, but AI systems generate outcomes that are often inscrutable, making actuarial assumptions tenuous. High‑profile mishaps—such as a generative model falsely accusing a solar firm of legal trouble and a deep‑fake fraud that siphoned millions—have already translated into multi‑digit payouts, prompting carriers like AIG and WR Berkley to seek regulatory relief. This trend underscores the difficulty of quantifying AI‑driven risk, especially when proprietary models evolve faster than policy language can keep pace.

Beyond isolated incidents, the real threat lies in systemic cascades. An agentic AI platform deployed across dozens of enterprises could malfunction in a synchronized manner, flooding insurers with thousands of claims in a single event. Existing commercial policies are calibrated for singular, catastrophic losses—typically up to $400 million per insured—but lack the bandwidth to absorb a deluge of smaller, concurrent claims. Regulators therefore face a dual challenge: they must balance the need for innovation with safeguards that prevent a domino effect capable of destabilizing the broader financial system. Potential solutions include industry‑wide AI risk pools, mandatory model transparency standards, and the development of bespoke cyber‑AI insurance products.

For businesses, the emerging coverage vacuum translates into heightened operational risk and potential cost overruns. Companies may need to internalize AI risk management, invest in explainable AI tools, and allocate capital for self‑insurance reserves. Meanwhile, insurers are likely to craft new policy endorsements that price AI exposure based on model provenance, usage scope, and governance controls. As the market adapts, firms that proactively address AI liability—through robust governance and strategic insurance partnerships—will gain a competitive edge, while laggards risk facing unmitigated financial fallout. The evolving landscape signals a critical inflection point where technology, risk, and regulation intersect.