%20(1).png)
Artificial intelligence is becoming part of routine business operations faster than insurance coverage is adapting. When those tools produce incorrect, biased, or harmful results, the question is not only who made the error. It is also whether the resulting loss fits within existing insurance coverage or is excluded altogether.
Gallagher's 2026 AI Adoption and Risk Survey points to rapid adoption, growing concern over AI errors, and early claims activity. 1
Gallagher's 2026 AI Adoption & Risk Survey
| Issue | Statistic | Why It Matters |
|---|---|---|
| Legal and reputational risk | 56% cited legal and reputational risks from AI misuse | AI use can be connected directly to potential liability and brand damage |
| Data protection and privacy | 55% cited data protection and privacy violations | Shows the overlap between AI risk and cyber/privacy exposure |
| AI-related claims activity | 20% said a client experienced an AI-related loss or claim in the past year | AI-related insurance issues are no longer theoretical |
| Active AI-related legal cases | Over 200 active legal cases involving AI and machine learning across cyber, EPL, product liability, and E&O 2 | AI liability is developing across several insurance lines |
Cyber Insurance as a Preview, Not a Playbook
Early cyber losses were often tendered under traditional property, crime, general liability, and professional liability policies before dedicated cyber products, exclusions, underwriting questions, and loss-control expectations emerged. AI coverage may follow a similar path.
The point is not that AI risk belongs in cyber policies, but that emerging technology risks often begin as coverage questions before the market develops clearer products, exclusions, endorsements, and underwriting practices.2 The American Bar Association has described "silent AI" as a developing coverage challenge, where AI exposures are neither expressly included nor excluded under traditional liability policies.³
AI risk may show up inside familiar claim types rather than as a clearly labeled "AI claim." The examples in the following table show how different AI uses can create different coverage, claims-handling, and defense issues.
AI-Related Claims: Use Cases, Coverage Lines, and Real-World Examples 4-10
| AI Incident | Alleged Harm | Possible Coverage Lines | Illustrative Case | Claim / Defense Issue |
|---|---|---|---|---|
| AI used to screen applicants or employees | Discrimination, bias, improper employment decision | EPLI | Mobley v. Workday alleged that AI-based applicant screening discriminated based on race, age, and disability. | Was the tool approved, tested, supervised, and documented? |
| AI drafts professional advice, reports, or client-facing work | Incorrect guidance, client loss, professional error | E&O / Professional Liability | Sullivan & Cromwell apologized to a federal bankruptcy judge after a court filing reportedly included AI-generated inaccuracies, fabricated citations, and misstatements. | Was there human review before the output was relied on? |
| Confidential data entered into an AI tool | Privacy breach, data misuse, regulatory exposure | Cyber / Privacy | OpenAI and Microsoft have faced privacy litigation alleging misuse of personal data to train AI systems. | Were data controls, vendor terms, and usage restrictions in place? |
| AI generates marketing, media, or public-facing content | Copyright, defamation, false statement, reputational harm | Media Liability / CGL | CNN v. Perplexity and other publisher suits allege AI tools unlawfully copied or distributed protected media content. | Who reviewed the content before publication or use? |
| AI embedded in product, service, or operational decision-making | Bodily injury, financial loss, product or service failure | Product Liability / CGL / E&O | Tesla Autopilot litigation has produced product-liability and negligence claims involving AI-enabled vehicle systems. Moffatt v. Air Canada held the company responsible for misinformation provided by its chatbot. | Was the AI output traceable, tested, and subject to escalation controls? |
| Board or management fails to oversee AI risk | Governance failure and inadequate controls | D&O | SEC “AI washing” enforcement actions against Delphia and Global Predictions show how AI claims can create governance, disclosure, and management-liability exposure. | Did leadership identify, monitor, and govern AI-related exposure? |
These examples are illustrative of emerging liability and claims issues; they do not determine whether any particular policy would respond.
How Alan Gray Can Help
Alan Gray can help clients prepare for a more developed AI coverage market by assessing claim-handling, governance, documentation, escalation controls and legal spend.
Clients that identify AI-related exposure, document controls, and track claim activity will be better positioned to respond to underwriting questions, coverage disputes, and litigation strategy.
Key Takeaway
AI coverage may develop like cyber coverage: uncertainty first, followed by claims experience, policy wording, exclusions, and underwriting discipline. The difference is that AI risk may spread across more coverage lines at once. For insurers and policyholders, the priority is to identify AI use, track related claims, and document controls that support coverage, claims handling, and litigation strategy.
Citations
- Gallagher. "The 2026 AI Adoption and Risk Survey: AI in Action." Arthur J. Gallagher & Co., 2026.
- Farley, John, and Dan Burke. 2026 Cyber Insurance Market Outlook. Arthur J. Gallagher & Co., 2026.
- American Bar Association. "The Evolving Landscape of AI Insurance: Empirical Insights, Risks, and Policy Gaps." The Brief, 5 Jan. 2026.
- Akin Gump Strauss Hauer & Feld LLP. "Court Allows Discrimination Claims Against AI Hiring Tool to Proceed \| Mobley v. Workday, Inc." AI Law and Regulation Tracker, 29 Apr. 2026.
- Freifeld, Karen, and Mike Scarcella. "Sullivan & Cromwell Law Firm Apologizes for AI 'Hallucinations' in Court Filing." Reuters, 21 Apr. 2026.
- Brittain, Blake. "OpenAI, Microsoft Hit with New U.S. Consumer Privacy Class Action." Reuters, 6 Sept. 2023.
- Stempel, Jonathan. "CNN Files Lawsuit Against Perplexity Alleging Unlawful Content Distribution." Reuters, 28 May 2026.
- Stempel, Jonathan. "Tesla Ordered by Florida Jury to Pay $243 Million in Fatal Autopilot Crash." Reuters, 1 Aug. 2025.
- American Bar Association. "BC Tribunal Confirms Companies Remain Liable for Information Provided by AI Chatbot." Business Law Today, 29 Feb. 2024.
- U.S. Securities and Exchange Commission. "SEC Charges Two Investment Advisers with Making False and Misleading Statements About Their Use of Artificial Intelligence." SEC, 18 Mar. 2024.
