Anthropic warns some risks may be uninsurable. AI coverage is going to be an issue for every company

A sign about AI security incidents sits on a stand during a New York City Council committee hearing about the risks of artificial intelligence and how it should be regulated. The hearing took place at City Hall on October 5, 2026, in New York City.

Adam Gray | Getty Images

Anthropic's IPO filing describes many risks, including stolen intellectual property, security problems, lost data, climate change, AI agents acting on their own, and even a possible threat to humanity's existence. Hidden among these warnings are basic questions about insurance: What will insurance cover? Can companies even buy insurance to protect them from all the ways AI could go wrong? Anthropic could not answer these questions.

“Because our models may behave in ways that no user or developer intended, claims arising from such actions could result in substantial liability. Our insurance policies also may not provide sufficient coverage to adequately mitigate the potential liabilities we face, especially where any one incident, or a group of incidents, causes disproportionate harm,” the company's S-1 filing states. “In addition, with the scale of harm that could result from a catastrophic AI-related incident, no single insurer, and potentially the insurance industry in the aggregate, may have sufficient capacity or expertise to underwrite insurance commensurate with such exposure. Further, the insurance market for AI-specific risks is nascent and evolving, and we may be unable to obtain coverage for certain AI-related liabilities on commercially reasonable terms, or at all, leaving us exposed to potentially significant uninsured losses.”

These risks are very different from AI chatbots helping people book airline tickets or find the best deals on a backyard grill. But there have been more cases of AI chatbots behaving in unexpected ways, and some companies are watching closely — even blocking them. In some serious cases, chatbots have been accused of helping cause a death, including a case reported this week. These cases raise difficult insurance questions without easy answers. AI is also becoming part of almost every area of business, increasing the risk that companies could face claims or have to pay for harm. This has put the insurance industry on alert.

Alissa Lugo, a senior director analyst in Gartner's legal and compliance practice, says her research suggests that claims involving AI playing a role in a person's death could exceed 2,000 worldwide by the end of 2026. This includes cases where AI played some part in a death. She expects that number to keep growing.

Lugo says more companies are adding exclusions for AI to their insurance policies. In the end, insurers, with help from the courts, will decide what AI insurance covers and how that coverage works. Insurance for cyber risks, in particular, often provides little protection against claims involving physical injury.

Many experts, including Lugo, say the situation with AI insurance today is similar to the early days of cyber insurance.

“Right now, cyber is the policy doing the most to cover AI, but it only covers one kind of AI loss. Cyber responds when AI is involved in a data breach, a privacy violation, a system outage, or a deepfake wire fraud,” said Dennis Hendrickson of Hendrickson Insurance, an independent insurance brokerage in Sarasota, Florida.

Hendrickson said some cyber insurers have changed their policy wording to include security incidents caused by AI. “That's useful. But nearly every cyber policy excludes bodily injury. Cyber was built for data and dollars, not for people getting hurt. This is where there is a major problem, in my opinion,” he said. He called the lack of coverage for bodily injury or physical damage a “gaping hole” that needs to be addressed.

Gaps in AI insurance coverage

Kara Higginbotham, head of professional liability and cyber for U.S. national accounts at Zurich North America, says AI is creating new risks across different types of insurance. But insurers are handling these risks in very different ways, depending on the type of policy. “Every line of business is handling things a bit differently,” she said.

Some companies, including Zurich, have a set of approved AI tools that employees can use, while other tools are blocked. Higginbotham said insurance policies may need to include safety measures for AI as a normal part of coverage in the future. For now, however, insurance generally still has a gap when AI-related incidents cause physical injury or other physical harm.

“It is a gap because it is not being covered by default. A lot of carriers would be wary to write that into a policy today, because as carriers, we have to be able to underwrite risk,” Higginbotham said. She explained that these situations are so new that insurers do not yet understand them well enough. In some cases, general liability insurance could be expanded to cover them.

“Insurers are saying, 'We can't cover what you can't show us.' Most companies can't say where their employees are using AI tools. That's the first thing to fix,” said Henrietta Akpata, an executive fellow and visiting professional lecturer of management and AI at Marist University.

Akpata says this problem affects more than just large technology companies. A local business could face problems if its website chatbot gives a customer the wrong answer. An HR team using AI to sort job applications could also face risks. “Those are everyday uses, and they may no longer be covered,” Akpata says.

Experts disagree on whether AI insurance will become a common, separate product or be added to existing cyber insurance policies. Hendrickson expects insurance for financial losses caused by AI to become common and to be priced like cyber insurance. But he says bodily injury caused by AI will be harder to insure. He expects insurers to examine risks closely, set low limits on certain types of claims, and require a person to check what AI does before its output reaches a patient. Insurance experts say some businesses connected to health care may not be able to buy this coverage at all.

Jaymin Kim, who leads emerging technologies in Marsh's cyber practice, says the company has not seen many AI-related insurance claims so far. She believes most current cyber insurance policies can cover AI-related problems. “AI is not a peril. It's just a technology through which existing familiar categories of harms and losses might manifest,” said Kim. She added that making AI its own insurance category could be a very complicated task.

“We're increasingly seeing AI being embedded in everyday business activities, and so by virtue of that, it's becoming invisible and ubiquitous. Which means on the insurance side, you cannot practically isolate AI in the causal chain of loss,” Kim said. She explained that if AI is involved in a data privacy or security incident, the incident is still a data privacy or security problem. She also warned that making customers deal with several overlapping insurance policies could become a nightmare.

“Are we saying now we're going to have one CGL policy for everything that doesn't touch AI, and then a second CGL policy for everything that does touch AI?” Kim said.

The insurance industry has not yet seen an entirely new type of risk created by AI. But Kim acknowledged that this could change, saying, “maybe the operative word is 'yet.'”

What cyber insurance history says about the AI risk future

Vikram Desai, global cyber strategy, risk and architecture lead at Accenture, said that until recently, businesses were focused on finding ways to use AI. “Now, it's about what happens when it goes wrong. That's exposed how badly AI fits the policies companies already own,” Desai said.

In some ways, this is similar to the early days of cybersecurity. “Cyber risk sat silently inside general liability for years until insurers named it, excluded it, then sold it back. AI is on the same curve, but compressed,” Desai said. “The real exposure is the AI companies don't know they have. Exclusions can bite when AI sits anywhere in the causal chain, including inside a vendor's product,” he added.

Desai says insurance underwriters are already asking companies where they use AI and who is responsible for overseeing it.

Higginbotham said most AI-related claims today are not about major attacks that shut down computer networks. Instead, many involve intellectual property. These include claims against AI developers accused of taking content without permission to train their models, as well as claims that AI models were trained using copyrighted material. “We also have seen data misuse or privacy-related claims,” Higginbotham said.

One problem with cyber insurance is that it may not be easy to tell whether AI was used in a social engineering attack. For example, an attacker might send an email or a Microsoft Teams message pretending to be a colleague.

“We know that bad actors are using AI in more convincing ways to carry out attacks, and that is buoying cyber claims, accelerating them,” Higginbotham said. But she added that it is not always clear what part AI played in an attack. “So it will not be a separate category,” she said, adding that cyber insurance is designed to cover this kind of risk.

She said that if insurers see “a sustained change in claim development patterns,” they may need to charge more for coverage if the growing risk of losses creates problems for their finances. “But we are not there yet,” she added.

“Our underwriting will have to evolve and adapt to changes in the threat environment. We are currently working on partnering with law firms and tech vendors to help,” Higginbotham said. “We are not always looking to exclude the next risk — we want to help companies manage risk,” she said.

Companies that have not put basic safety measures in place may eventually face insurance exclusions if they refuse to improve how they manage risks. “But we don't jump to exclusions as the first arrow in the quiver,” Higginbotham said.

Like other experts, Hendrickson sees similarities between AI insurance and the early days of cyber insurance. He is waiting for the large wave of claims he believes is coming, and says no one is ready for it, including the biggest insurance companies.

“We've seen this before with cyber insurance. First insurers stopped covering it, then they created separate policies for it, and their checklists ended up setting the rules companies followed. I expect the same with AI: for many businesses, insurers will set the standards before the government does,” he said.

“Insurance companies come out with a new product, then the claims come in, and they won't renew the policy,” Hendrickson said. He added that independent insurance agents like him often have to deal with angry customers whose claims are not covered. “We're the ones in the trenches dealing with this every day.”