The Safety Net Is Shrinking: Insurers Move to Exclude AI Risks

As companies race to integrate artificial intelligence into their operations, a critical safety net might be quietly disappearing in the process.

Major insurance providers, including AIG and W.R. Berkley, are reportedly seeking regulatory permission to exclude AI liabilities from standard corporate policies. The industry is moving to limit its exposure to what it views as "unpredictable and opaque" technology, according to the Financial Times,

This shift represents a fundamental re-evaluation of risk that could have massive ripple effects on how, and if, enterprises deploy AI agents and automated tools.

I discussed this developing trend on Episode 183 of The Artificial Intelligence Show with SmarterX and Marketing AI Institute founder and CEO Paul Roetzer, who spent more than a decade working closely with the insurance industry. 

Are AI Hallucinations Too Risky to Cover?

Insurers are in the business of calculating risk, but generative AI is proving difficult to quantify.

One proposal from W.R. Berkley would reportedly bar claims involving any actual or alleged use of AI, including products sold by a company that merely incorporate the tools. Meanwhile, Chubb has agreed to cover some risks but specifically excludes "widespread" incidents where a single model failure affects many clients simultaneously; insurers fear this scenario could lead to systemic, aggregated losses.

These moves follow several high-profile, costly incidents:

• The Deepfake Scam: A deepfake video conference scam cost engineering group Arup $25 million.
• The Chatbot Error: A tribunal forced Air Canada to honor a refund policy invented by its customer service chatbot, which the airline had argued was a separate legal entity responsible for its own actions.
• Defamation Suits: Companies such as Wolf River Electric have sued tech giants for defamation after AI overviews hallucinated or made up non-existent legal troubles.

For insurers, these "hallucinations" and errors fall into a gray area that makes them too risky to underwrite under current standard liability or cyber policies.

An Overlooked Risk

For Roetzer, who owned a marketing agency for 16 years that worked extensively with insurance carriers and agent networks, this development highlights a blind spot for many business leaders.

“I spent a lot of time thinking about the insurance industry for well over a decade,” says Roetzer. “I honestly hadn't really stopped and thought deeply about the implications of AI on insurance policies. But now that I saw this topic, my mind is kind of racing.”

If insurers won't protect firms from AI risks, the internal demands for reliability will skyrocket. Companies may become "gun shy" about adopting AI if they know a single hallucination or agent error could result in an uninsured, multi-million dollar liability.

AI Agents Can Complicate Things

The timing of these exclusions is particularly notable as the industry pivots toward "agentic" AI, systems that can take autonomous actions, execute code, and make decisions without human intervention.

“There are definitely risks, especially as we start getting more and more into the agentic side of this, that I would imagine most businesses have not contemplated yet in relation to their insurance,” says Roetzer.

While a chatbot answering a question incorrectly is problematic, an autonomous agent executing financial transactions or modifying code creates a liability that standard business insurance is not designed to cover.

What Can You Do?

This trend is still in its early stages, but it is moving fast.

If you are a business leader, now is the time to review your contracts and speak with your risk management teams. The assumption that your general liability or errors and omissions (E&O) policy covers your new AI tools might no longer be true.

“If you're in the insurance space or if you deal with contracts for your company, this is something that's probably very near term for you,” says Roetzer.

As AI technology accelerates, companies must work hard to keep up and protect themselves.

“It’s still a pretty early trend, but definitely seems like this could have big ripple effects over time,” Roetzer says.