IBM Study Finds 80% of EMEA Enterprises Escalate AI Incidents to Board as Governance Struggles to Keep Pace

Image: Illustrative image · Viscovery, IBM, Amazon · Public domain · Source

IBM research shows eight in ten EMEA enterprises reported AI incidents to boards, yet only 40% updated AI governance, highlighting gaps in accountability and risk management amid growing AI adoption.

A new study conducted by IBM across Europe and the Middle East reveals that a vast majority of enterprises are facing significant AI-related incidents, with these issues frequently escalated to the highest levels of corporate leadership. According to IBM's 2026 research, 80% of surveyed executives reported at least one AI-related incident reaching their board or C-suite within the past year. These incidents include AI agent failures, data exposures, decision errors driven by AI, and security breaches.

Despite this high incidence and visibility, less than half (40%) of the organizations have updated their AI governance policies or frameworks in direct response. This suggests a concerning mismatch between the pace of AI deployment and the readiness of governance and compliance mechanisms.

Rather than pausing AI adoption until governance frameworks catch up, the majority (56%) of executives reported proceeding within agreed risk limits, with only a third delaying deployments. The rapid integration of increasingly autonomous AI agents, with 35% of respondents already running such systems in production and 75% expecting agent autonomy in decision-making by 2027, places additional strain on existing governance structures. Only 37% believe their current governance systems are capable of scaling alongside this growing autonomy.

The accountability landscape remains unclear, despite nearly all organizations (98%) having formal board-level AI risk ownership. Assigned responsibility for adverse outcomes from autonomous AI agents is distributed, with the CIO or CTO cited most frequently but by only 17% of respondents. Legal liability views vary: around 20% advocate proportional shared liability between deployer and vendor, 19% favor the establishment of regulatory frameworks, and 14% hold vendors fully liable.

To safely manage AI autonomy, executives anticipate the need for formal legal and contractual liability frameworks for vendors (33%), defined human escalation thresholds (32%), real-time monitoring with audit trails (30%), and emergency stop capabilities for AI agents (29%).

Additional findings indicate governance maturity gaps: only 37% have formal, board-reviewed quantifications of financial risk related to AI governance gaps. Supply chain risks are also a concern, with 54% reporting governance shortcomings in third-party AI systems and only a minority continuously monitoring third-party AI risks or applying independent assurance.

Regional regulatory environments influence governance investment priorities: EU respondents demonstrate proactive investment ahead of the EU AI Act’s December 2027 deadlines, while UK executives cite sector regulatory expectations, and those in Saudi Arabia and the UAE highlight national AI strategy mandates.

IBM’s study, conducted by Censuswide in September 2026 with over 3,200 senior executives from eight countries, underscores the urgent need for enterprises to enhance AI governance frameworks. Effective, scalable governance with clear accountability is crucial to realizing the full business value of AI while managing risks in an era of increasing AI autonomy.