The Lloyd’s Market Association (LMA) has released findings from a survey examining potential artificial intelligence-related losses across four insurance sectors, offering insight into underwriters’ assessments of emerging risks as AI adoption accelerates.
The survey, titled Understanding AI Exposures: AI Loss Scenarios Survey Results, gathered responses from 144 managing agents in mid-2025, with 94% coming from underwriters. Participants assessed the likelihood, potential severity, and overall impact of specific AI loss scenarios in professional indemnity, cyber, product recall, and accident and health insurance.
Professional indemnity emerged as the highest potential impact area, according to the findings. Underwriters pointed to the widespread use of AI tools in professional services and the risk of substantial losses when erroneous advice or analysis causes harm.
In cyber insurance, computer system downtime caused by AI malfunction was seen as plausible. However, underwriters said confidence in insureds’ risk management practices moderated their overall impact assessments.
The product recall scenario was considered highly unlikely at present because of the limited current use of AI in design and manufacturing processes. However, underwriters expect adoption to increase in these areas over time.
For accident and health insurance, the self-driving vehicle scenario was assessed as plausible, with mixed views on the potential severity of losses.
“Use of AI software is accelerating across every sector and underwriters are already seeing early indicators of how these technologies could contribute to future claims,” said David Powell, head of technical underwriting at the LMA.
“This opinion survey demonstrates that, while current levels of concern remain moderate overall, there is a clear expectation that exposures will grow as AI becomes more deeply embedded in operational and decision-making processes.”
The survey asked respondents to assess the viability, potential magnitude, and overall impact of specified AI loss scenarios, while also providing views on insureds’ current use of AI and associated risk management practices.
Powell noted: “The findings will help inform further market discussion on how best to assess, manage, and underwrite these emerging risks.”
