IBM survey exposes gaps in business AI oversight

An IBM survey of over 3,000 business leaders across EMEA found that enthusiasm for expanding AI deployments is outpacing safeguards. Among respondents, 80% said an AI-related incident had reached their board or C-suite over the past year. Those incidents included agent failures, data exposure, decision errors and security breaches. Despite that, fewer than half (40%) of those organisations said they would change their AI governance policies or frameworks in response.

The gap could become more consequential as agents gain decision-making authority. The survey found 75% of respondents expected AI agents to have some level of decision-making autonomy by 2027, while only 37% believed their governance was designed to scale with agentic AI deployments. Formal oversight is widespread: 98% of executives surveyed had assigned responsibility for AI risk at board level, including 26% with a dedicated board-level AI committee.

Leaders remained divided over legal liability when agents act independently. Measures being developed included frameworks defining vendor liability for agent failures (33%), human escalation thresholds (32%), real-time monitoring and audit trails (30%), and enterprise-wide controls to halt agent activity immediately (29%). IBM executive Ana Paula Assis said governance must become an operational business requirement, with clear boundaries, ownership and real-time controls rather than remaining a compliance exercise.