8 in 10 executives say an AI incident reached the board
Only 40% of those organizations changed their AI governance policy after the incident.
On 14 October 2026, IBM published a survey of 3,221 senior executives in Europe and the Middle East. 8 in 10 say an AI incident reached their board or top management during the last year. Only 40% of those organizations changed their AI governance policy after it. IBM says governance does not keep pace with the use of AI, and AI agents will test it more.
Agents in production
35% of the executives operate AI agents in production at this time. Only 37% say their governance system can scale with agents. Most think agents will make some decisions without a person by 2027.
Who is responsible
The survey found no clear agreement on who is responsible when an agent makes an error. The CIO or CTO is the role that executives chose most frequently, but only 17% chose it.
By 2027, the executives think they must have these controls:
- legal or contract rules for the liability of the vendor for agent failures
- set limits for when a human must start to control the work
- real-time monitoring and audit trails of agent decisions
- a control to stop all agent work at one time
Vendor risk
54% say their governance has gaps for the AI systems of vendors and partners. Only 21% monitor third-party AI risk at all times.
IBM view
Ana Paula Assis, a senior vice president at IBM, says governance must be part of how agents operate. Assis says each agent must have clear limits and a clear person who is responsible when it makes an error.
Sources
Posted
