Insurers zero in on AI disclosures as D&O lawsuits spread beyond tech

D&O in the AI era: Insurers zero in on corporate disclosures

Artificial intelligence is now central to strategy-and to securities litigation. Regulators and plaintiffs' firms are combing through filings, earnings calls, marketing, and product docs for AI statements that overpromise or omit material limits. The result: a rapid rise in disclosure cases, with non-tech companies firmly in scope.

D&O underwriters are responding by scrutinizing how insureds talk about AI and adjacent technologies. Policy wording hasn't broadly shifted, but the underwriting interview has. As one broker put it, "Underwriters are looking at disclosures and what companies are representing to the market to their investors."

Litigation is up-and it's not just tech

AI-related securities suits accelerated in the first half of the year, with at least a dozen cases-nearly matching the 15 filed in all of 2024. This isn't confined to software vendors. AI shows up in hiring, pricing, underwriting, logistics, safety systems, and customer support across sectors.

A retailer using demand forecasting tools or a hospital piloting triage algorithms can face the same disclosure risk as a pure-play tech firm if public statements stretch claims or skip constraints. "Anybody that's leveraging AI is potentially at risk," said Rik Goyton, executive vice president and financial lines practice leader at CAC Specialty.

Echoes of past cycles

Disclosure-driven securities claims are familiar territory. The themes change-ESG and greenwashing yesterday, Y2K before that-but the mechanics stay constant: statements to the market that don't line up with reality. Plaintiffs keep looking for the next angle; AI is the latest.

Underwriting focus: your AI narrative and the evidence behind it

Insurers are pressing for proof that AI narratives match actual deployment and controls. The concern shows up in Q&A and diligence, even if exclusions haven't moved. Expect questions about governance, validation, and how limitations are communicated to investors.

• Inventory and ownership: What AI systems are in use (internal, vendor, open-source, LLMs), where they sit, and who is accountable.
• Model risk controls: Validation, testing, monitoring, and change management. Evidence, not promises.
• Bias and fairness: Screening for disparate impact in hiring, underwriting, pricing, and safety applications; remediation workflow.
• Disclosure controls: How AI claims in 10-Ks/10-Qs, earnings scripts, marketing, and product docs are harmonized and approved.
• Metrics and limitations: Are KPIs reproducible? Do you state known constraints, data dependencies, and error rates?
• Third-party risk: Contract terms, audit rights, indemnities, and oversight for vendors providing AI.
• Board oversight: Reporting cadence, risk appetite, and escalation paths for AI incidents or material changes.

Claims handling: slower, higher scrutiny-especially in excess

Carrier profitability pressures are shaping claim resolution. Expect longer timelines, more internal approvals, and increased use of outside counsel, particularly on mid-to-high excess layers. "They're approaching the process with much more diligence... Many [claims] are being elevated to higher levels of authority and approval," Goyton noted.

For insureds, that can translate to more people on calls and more touchpoints before a deal gets done. Effective broker advocacy can buffer the friction, but plan for a longer runway.

Pricing stays buyer-friendly (for now)

D&O rates for 2025 renewals are generally flat to down about 5%, with decreases slowing. Favorable pricing doesn't offset disclosure risk, though. The better your controls and narrative discipline, the less drama at renewal and in a claim.

IPO/SPAC, crypto, and EPL remain active

As IPOs and de-SPACs return, expect more suits tied to prospectus statements and strict-liability exposures under the Securities Act. Crypto-related filings continue. On the EPL side, AI-driven hiring and interview tools create additional plaintiff angles if outputs are biased or claims about accuracy slip into puffery.

AI-washing is a red flag-inside and outside insurance

• Don't overstate: Avoid absolute claims (e.g., "eliminates bias," "guarantees accuracy," "fully autonomous").
• Disclose limits: Data requirements, error rates, human-in-the-loop, and operational constraints.
• Keep receipts: Retain testing, validation, and monitoring evidence that supports public statements.
• Align teams: Investor relations, legal, product, marketing, and risk must use the same language and KPIs.
• Update quickly: If performance shifts or a dependency breaks, revise disclosures before the next call or filing.

Regulators are already signaling expectations. See the SEC's actions on "AI-washing" in the investment space for tone and direction of travel: SEC charges on misleading AI claims. The FTC has similar guidance on keeping AI marketing claims in bounds: Keep your AI claims in check.

What to prepare before your next D&O renewal

• One-page AI overview: Where AI is used, objectives, and risk controls-matched to your disclosures.
• Evidence pack: Validation summaries, fairness testing highlights, monitoring dashboards, and incident logs.
• Disclosure map: Crosswalk of AI statements across SEC filings, earnings scripts, marketing, and product docs.
• Governance exhibit: Policies, RACI, board reporting, and third-party oversight.
• Playbooks: Escalation for model drift, vendor failure, or material misstatement; PR and investor comms templates.
• Securities counsel check: Pre-clear key AI claims in 10-K/10-Q and earnings materials for consistency and accuracy.

Bottom line for insurance professionals

The exposure is less about the algorithm and more about what you say it does. Tighten the narrative, prove the controls, and keep filings, earnings scripts, and marketing in sync. Do that, and you'll undercut plaintiff theories while giving underwriters confidence to keep pricing stable.

If your team needs structured upskilling to standardize AI practices and documentation, consider focused training resources like AI certifications that cover governance, validation, and disclosure fundamentals.

Key quotes

• "Underwriters are looking at disclosures and what companies are representing to the market to their investors."
• "Anybody that's leveraging AI is potentially at risk."
• "They're approaching the process with much more diligence... Many are being elevated to higher levels of authority and approval."

Immediate actions this quarter

• Run a redline across your 10-K/10-Q, earnings scripts, and marketing for AI claim consistency.
• Remove absolute claims and add plain-English limitations where appropriate.
• Prepare a concise AI governance and validation brief for underwriters.
• Align IR, legal, and product on a single AI narrative-and stick to it.

Keep it simple: say what the tech does, prove it, and disclose what it can't do. That's your best D&O defense in the AI cycle.