Major insurance companies are sounding the alarm: artificial intelligence is becoming too unpredictable and dangerous to cover. Industry leaders like AIG, Great American, and WR Berkley are actively seeking regulatory approval to exclude AI-related liabilities from their corporate policies, signaling a growing crisis in how businesses manage AI risk.
The “Black Box” Problem
The core issue is the opacity of many AI models. Underwriters describe them as “black boxes,” meaning their decision-making processes are difficult or impossible to audit. This creates a massive liability gap because insurers cannot accurately assess the potential for errors or malicious use. The financial consequences of such errors are already materializing:
- Google’s AI Overview scandal: A false accusation generated by Google’s AI led to a $110 million lawsuit against a solar company in March.
- Air Canada’s chatbot blunder: The airline was forced to honor fabricated discounts offered by its own chatbot.
- Arup fraud: A digitally cloned executive was used in a deepfake video call to steal $25 million from a London-based firm.
Systemic Risk: The Real Threat
Insurers aren’t merely worried about isolated incidents; they fear systemic collapse. A single AI failure impacting thousands of companies simultaneously could trigger catastrophic payouts. An Aon executive explained that while a $400 million loss is manageable, a widespread “agentic AI mishap” causing 10,000 simultaneous claims would bankrupt even the largest insurers.
This uninsurability creates a perverse incentive: businesses may be less careful with AI if they know insurers won’t cover the fallout. The lack of financial backstops could slow AI adoption or force companies to internalize risks they are not equipped to handle.
The Future of AI Risk
The situation underscores a critical gap in the current AI landscape. As AI becomes more powerful and autonomous, it also becomes more unpredictable. Until better safeguards and transparency are implemented, insurers will remain hesitant to cover AI-related liabilities, leaving businesses vulnerable to potentially ruinous losses.
The exclusion of AI risks from insurance policies is a clear signal: the industry sees the dangers as too great to ignore. This may be the first step toward a more cautious, regulated approach to AI deployment in the corporate world.
