IBM survey finds 80% of EMEA leaders faced AI incidents at board level last year

An IBM report based on a survey of more than 3,000 business leaders across the EMEA region indicates that AI-related incidents have become a frequent boardroom concern. According to the findings published by TechRadar, eighty percent of respondents stated that an event such as an agent failure, data exposure, or security breach reached their executive leadership within the last year.

Despite the high frequency of these escalations, organizational responses remain limited. The data shows that fewer than half of the surveyed entities, specifically forty percent, intend to modify their AI governance policies or frameworks. This discrepancy highlights a growing gap between the operational reality of AI risks and the strategic adjustments businesses are willing to make to manage them.

Governance gaps and future autonomy

The survey highlights a significant disconnect between expected AI capabilities and current governance readiness. Three-quarters of the respondents anticipate that AI agents will possess some level of decision-making autonomy by 2027. However, only thirty-seven percent believe their existing governance structures are designed to scale alongside agentic AI deployment. This suggests that most organizations are unprepared for the increased complexity that autonomous systems will introduce to their operations.

Ana Paula Assis, IBM Senior Vice President and Chair for EMEA and APAC, noted that the industry conversation has shifted from simple adoption to governance and accountability. She emphasized that as organizations move from AI assistants to autonomous agents, governance must be treated as a business requirement rather than a mere compliance exercise. She argued that enterprises scaling AI with trust and accountability will realize the greatest business value.

Board responsibility and liability

Awareness of AI risks appears high at the executive level, with ninety-eight percent of surveyed leaders formally assigning AI risk responsibility to the board. Twenty-six percent of these organizations have established a dedicated board-level AI committee to oversee these issues. Despite this structural awareness, the survey reveals deep divisions regarding legal liability when AI agents make independent decisions.

Respondents identified several key areas where frameworks need strengthening. Thirty-three percent cited the need for formal legal or contractual frameworks defining vendor liability for agent failures. Other priorities included defining human escalation thresholds at thirty-two percent, implementing real-time monitoring and audit trails for thirty percent, and establishing the ability to immediately halt agent activity across the enterprise for twenty-nine percent.

What should you do?

Organizations should review their current AI governance policies to ensure they address autonomous decision-making and agent failures. Leaders need to verify that board-level oversight includes clear definitions of liability for both internal deployments and third-party vendors. Companies should also assess whether their monitoring tools can provide real-time audit trails for AI actions. Finally, businesses must establish clear protocols for halting agent activity if unauthorized or harmful decisions occur.

What to do and how to stay safe: IBM

• Review board-level governance policies to ensure they explicitly cover autonomous AI agent decisions and potential failures.
• Verify that contracts with AI vendors include clear liability clauses for agent errors or security breaches.
• Implement real-time monitoring and audit trails to track AI agent actions and decisions within your environment.
• Establish and test protocols for immediately halting AI agent activity across the enterprise in case of anomalies.

Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation

General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

How many business leaders reported AI incidents to their board last year?

According to the IBM survey, eighty percent of the more than 3,000 EMEA business leaders surveyed reported that an AI-related incident reached their board or C-suite.

What percentage of organizations plan to update their AI governance?

Only forty percent of the surveyed organizations stated they would be changing their AI governance policy or frameworks in response to recent incidents.

What do leaders expect regarding AI autonomy by 2027?

The survey found that seventy-five percent of respondents expect AI agents to have some level of autonomy in making decisions by 2027.