In July, the Government published its AI strategy for insurance. This is an ambitious task: adopting new technology is remarkably difficult. Research from Boston Consulting Group suggests that some 70% of technology projects fail to meet their objectives, and yet only a small proportion of AI investment (3 to 6%, according to analysis of data produced by Deloitte) is directed towards employee training and AI literacy.

At a recent Institute of Risk Management roundtable with a delegation from the British Malaysian Chamber of Commerce, we discussed human capital and people risk, followed by cyber and AI. During the conversation, I suggested that, increasingly, these are not separate subjects at all. AI risk is people risk.

Geopolitical tension in the Middle East has rarely remained localised and has usually had implications beyond the region itself. That is also true for the insurance market, where events rarely remain contained within national borders. For insurers, reinsurers and policyholders, instability in the region can quickly translate into questions around pricing, coverage, claims exposure, sanctions compliance and the availability of reinsurance support.

Mosaic Insurance has launched direct political violence underwriting from its European headquarters in Cologne, Germany, offering up to €100m (£85m) of syndicated capacity to European brokers.

Organisations are recognising workplace violence and active assailant incidents as major security risks, but many lack confidence in their ability to respond effectively, according to research from ASIS International. The 2026 Active Assailant Preparedness Report found that workplace violence and active shooter incidents ranked as respondents’ top two security concerns.

Cyber risk intelligence provider Kynd has launched a new AI detection capability that allows cyber insurers to identify AI technologies operating across an organisation's external digital footprint before assessing the risk.