AI Losses Are Hiding Inside Other Insurance Claims, Allianz Says

Estimated reading time: 8 minutes

Businesses now see AI as the second biggest business risk, according to Allianz’s annual ranking. But in insurance claims data, AI does not even make the top 20. Allianz Commercial explains that most AI losses are hidden under other categories. Their new report breaks these losses into four types and tracks the growing number of lawsuits.

Businesses list artificial intelligence (AI) as one of their main concerns. However, only a few insurance claims are labeled as AI-related. Allianz Commercial says this gap is due to the way claims are categorized.

Confirmed AI-related losses remain relatively low, the insurer’s new report says. AI losses are “frequently underreported, difficult to quantify or attribute.” Many land under cyber, directors’ and officers’ (D&O) liability, or product liability. Most insurers are not yet set up to track AI-driven claims, the report adds.

Allianz Commercial has set up an AI Risk and Claims Center of Excellence. Michael Daum, who leads the center, says confirmed AI losses are “almost certainly just the tip of the iceberg.”

Erin Kenneally made a similar point on the CINI podcast in 2025. She said AI is so deeply embedded that it is hard to identify its role in a loss. A year later, she told CINI that her warning still holds true.

Why AI Claims Are Hard To See

The report says AI use is often hidden or not documented. Employees may use public AI tools or built-in software features without official approval. This makes it hard to trace AI’s role in a loss. The categories can also overlap. For example, a facial recognition tool that violates privacy could be considered an AI claim, a privacy claim, or both. Similarly, a D&O loss involving misleading AI statements raises the same issue.

Allianz is creating a framework to track AI claims based on the cause of loss. The goal is for this framework to cover all lines of business. CyberCube used a similar method this year by mapping types of AI events to policy types and letting each insurer fill in the details.

Four Types Of AI Loss

The report sorts AI losses into four baskets:

  • Direct AI loss: This happens when an AI system causes harm through a mistake, decision, or output. For example, a truck crash victim claimed that a driver-assist system failed to detect a vehicle in its blind spot.
  • AI-related loss: In these cases, AI is a major factor. The loss results from governance, legal, or third-party issues involving AI. The main example is lawsuits over AI washing.
  • AI-accelerated loss: Here, AI increases the scale of harmful activities, like deepfake fraud or automated cyberattacks.
  • AI-dependency loss: In this case, AI does not cause the event, but heavy reliance on AI and poor backup plans make the loss worse or last longer.
See also  Cyber Risk Strategy Shifts as Directors Face Increasing Cyber Threats: Willis Survey 2025

Allianz is already monitoring AI claims that are worth millions of euros. The company says that cases like the truck crash or an AI washing lawsuit could each cost the industry about $100 million.

AI Litigation Is Climbing Every Year

The number of reported AI incidents continues to grow. According to the report, the AI Risk Initiative at the Massachusetts Institute of Technology (MIT) counted 179 incidents in 2023. This increased to 307 in 2024 and 453 in 2025.

[CHART VIDEO: ai-incidents-claims-allianz-2026.mp4]

Bar chart of reported AI incidents tracked by the MIT AI Risk Initiative, rising from 179 in 2023 to 307 in 2024 and 453 in 2025

The report, citing Cornerstone Research, says AI-related securities class actions increased from seven in 2023 to 16 in 2025. According to The D&O Diary, investors filed 22 such cases in the first eight months of 2026. This means 2026 could nearly double the total from 2025.

By mid-2026, artists and media companies had filed 115 intellectual property lawsuits against AI developers. Some law firms have also faced court sanctions for submitting filings with AI-generated citations.

“AI output should not be taken as gospel,” says David Ackerman, Allianz Commercial’s head of global financial lines, claims.

D&O Professionals Expect More AI Claims

Allianz and The D&O Diary surveyed 250 industry professionals across 25 countries. Among them, 86% expect AI-related claims to have a major impact on D&O insurance. In addition, 92% believe that poor AI governance increases the chances of D&O claims.

AI washing and data privacy were tied as the top risks, with each ranked first by 28% of respondents. Intellectual property disputes were next, at 20%.

Dan Holloway leads professional indemnity and management liability in Allianz Commercial’s financial lines. “AI is as much a governance issue as it is a technology issue,” Holloway says.

AI-Accelerated Attacks And Fraud

The report says criminals are already using AI to create more convincing phishing messages, produce deepfakes, and automate attacks. According to the Federal Bureau of Investigation (FBI), AI-related cybercrime cost Americans $893 million in 2025. CrowdStrike also reported an 89% increase in attacks by AI-enabled adversaries.

In 2024, the engineering firm Arup lost £20 million when a deepfake video call tricked an employee in Hong Kong.

See also  One AI Failure, Four Policies: AXA XL Names The Scenarios

On the CINI podcast, Gambit Security’s Jake Bernardes said AI lets one attacker look like a nation-state.

Dependency Turns An Outage Into A Bigger AI Loss

Allianz managed a case where a cyber-related outage affected a large retailer’s supply chain platform. AI was not the cause of the outage, but the retailer used AI for forecasting, inventory, and warehouse operations on that platform. When the system failed, distribution centers had to switch to paper processes, resulting in incorrect product picks, inventory problems, and delays.

In another case, a cyberattack forced a large company’s staff to process millions of dollars in transactions manually, using pen and paper.

Daum says that dependency on AI contributed more to these losses than the original outage or attack. Weak backup plans made the disruption last longer and increased the costs.

One AI Model, Many Losses

The report also notes that AI brings systemic risk. Many systems are reliant on the same models, training data, cloud infrastructure, and data centers.

“An issue with one model could become a worldwide issue,” says Michael Bruch. Bruch is Allianz Commercial’s global head of risk consulting advisory services.

AI systems operate based on probabilities. Even the best-designed systems can make mistakes. When used at scale, a small error rate can lead to thousands of incidents at the same time.

The report also mentions AI lab tests where models escaped test environments and hacked into live systems. CINI discussed these incidents and the debate about who is responsible for the costs.

What Allianz Wants From AI Claims Data

The report says AI can also help reduce losses. AI-powered detection can lower false positives by up to 70% compared to older systems. Early detection can prevent a €20,000 incident from turning into a €20 million loss.

Allianz Commercial sells the insurance and risk consulting the report recommends.

Daum sees current AI claims as an early warning. He says the number of claims is still low because companies are careful with AI and keep humans in control.

Daum believes the solution is to create an industry standard for defining and categorizing AI claims. The report says that without a clear way to collect AI loss data, the true impact of AI losses remains hidden.

FAQ – AI Insurance Claims

Why are AI insurance claims hard to track?

AI losses often land under cyber, D&O or product liability. Staff also use AI tools without approval, which hides AI’s role in a loss. Allianz says most insurers are not yet set up to track AI-driven claims.

What are the four types of AI loss?

Allianz Commercial names direct AI loss, AI-related loss, AI-accelerated loss and AI-dependency loss. They run from an AI system causing harm itself to heavy reliance on AI making a loss worse.

What is AI washing and why does it matter for D&O insurance?

AI washing is overstating what a company’s AI can do. Investors sue when the claims fall short. In an Allianz survey, D&O professionals ranked it the top AI exposure, tied with data privacy.

How fast is AI litigation growing?

AI-related US securities class actions rose from seven in 2023 to 16 in 2025. Investors filed 22 by the end of August 2026, Allianz reports.

How can AI dependency make a loss bigger?

When an outage takes down AI-driven processes, companies without fallbacks fall back to manual work. Allianz cites a retailer whose distribution centers reverted to paper, causing errors and delays.

What should boards do about AI risk?

Find where AI is used and how it shapes decisions. Then decide which AI risks to mitigate, accept or transfer. Keep manual fallbacks for critical processes.

Leave a Comment

×