UK firms embrace AI but remain divided over who owns the risk
An IBM survey of more than 3,000 business leaders across EMEA found that 80% had seen an AI-related incident reach their board or C-suite in the past year. Only 40% said they would change their AI governance policies in response. While 75% expect AI agents to have some decision-making autonomy by 2027, just 37% believe their governance is ready to scale with that shift.
IBM’s survey covered over 3,000 EMEA business leaders. Four in five reported an AI incident—such as agent failure, data exposure, decision error, or security breach—reached board or C-suite level in the prior year. Yet only 40% planned governance policy changes in response.
Almost all respondents (98%) said AI risk responsibility is formally assigned at board level, and 26% have a dedicated board AI committee. Still, liability views remain divided. Common steps include vendor liability contracts (33%), human escalation thresholds (32%), real-time monitoring (30%), and enterprise-wide agent shutdown capability (29%).
If governance lags adoption, employees, customers, and organizations may face greater exposure to automated decisions, data leaks, or unclear accountability. Boards and executives could be drawn into disputes over liability, while workers may need clearer escalation paths and oversight. The split on responsibility may slow trust in AI agents, affecting how quickly firms deploy them in everyday work. Regulators and vendors may also face pressure to clarify roles.