Clearspeed: AI adoption exposes insurance 'verification gap'
September 03, 2026 at 12:00 PM
As insurers accelerate the use of artificial intelligence in claims, underwriting and customer interactions, they may be creating a new vulnerability: Automated systems can process information faster than carriers can determine whether that information is trustworthy.
That is the central warning in The Speed of Trust: Building the Trust Intelligence Layer for Insurance in the Age of Agentic AI, a new report commissioned by Clearspeed and independently authored by insurance innovation strategist Sabine VanderLinden, CEO of Alchemy Crew Ventures. The research argues that insurers need to treat verification as core infrastructure rather than a downstream fraud-control function.
The analysis reviewed 76 public filings from 49 insurers and reinsurers, 31 industry studies, and interviews with 16 claims and underwriting leaders in the U.S. and U.K. Researchers found no mentions of "synthetic media," "synthetic identity" or "voice cloning" in the filings. Only six of the 49 companies mentioned deepfakes, according to the research, and those references addressed cybersecurity rather than potentially manipulated evidence entering claims or underwriting workflows.
The omission comes as insurance professionals acknowledge a growing threat. Verisk research cited in the report found 98% of U.S. insurance professionals surveyed said AI editing tools are fueling increased digital media fraud, but only 32% were very confident they could identify a deepfake.
"That is the verification gap: The distance between what the industry can see coming and what it can currently detect," VanderLinden said in a prepared statement. "Insurance is automating decisions faster than it can verify the information behind them."
For insurance professionals, the concern extends beyond fraud detection. AI agents can increasingly file claims, attach evidence and support automated decisions, but faster processing can also allow false photos, documents, voices or other information to move through systems with less human scrutiny. The report warns that automation can therefore magnify the consequences of unreliable inputs.
At the same time, overly aggressive verification can create friction for legitimate customers. Fraud is estimated to affect about 10% of property-casualty losses, meaning the overwhelming majority of claims are legitimate.
Nine out of ten customers "who make a claim are honest, good people for whom we, as insurance professionals, should be just sorting out their service needs as quickly as possible," former Zurich Insurance Group Chief Claims Officer Ian Thompson told researchers.
The report recommends creating what it calls a "Trust Intelligence Layer" across application, underwriting, renewal, first notice of loss, claim review and settlement. The concept uses a continuous risk signal to distinguish interactions that can proceed quickly from those requiring closer human review. Importantly, the signal informs rather than makes the decision.
For claims and underwriting leaders, the takeaway is that AI governance increasingly needs to address not only whether automated systems perform correctly, but whether the evidence and information feeding those systems can be trusted. As the report puts it: "As the automation rises, the assurance has to rise with it."
Editor's Note: We used artificial intelligence to draft this story. All AI output was reviewed by Editor-in-Chief Elana Ashanti Jefferson.
(Featured image credit: Kiattisak via Adobe Stock)
Read article here.
