Survey: most insurers want AI limited to insurance-specific models

83% would hand repeatable work to AI, but 75% insist on an insurance-specific or governed model, mea Platform found.

New research from mea Platform found 83% of respondents in the global insurance market would hand repeatable operational work to AI, and 75% would only let AI execute that work through an insurance-specific or governed model. Just 6% said they would trust a generalized AI model without insurance-specific governance. The release did not disclose sample size or how respondents were selected.

The source's own statement, as linked above. Nothing beyond it has been independently checked by us.

Sample size not disclosed, so these percentages describe stated attitudes among an unspecified group, not a measured or representative outcome for US carriers specifically.

If a vendor pitches a generalized model for underwriting or claims work, this survey is a reason to ask about its insurance-specific governance, not proof that generalized models are unsafe. Ask vendors for their model's training data scope and governance controls in writing.

Acrisure unveils Auris, an AI platform across its operations