Intelligent agent incidents breach boards of major UAE companies: accountability crisis precedes regulation

AI failures have reached closed decision-making rooms in three out of four major organizations in the UAE, placing senior executives and boards of directors before a unprecedented dilemma in pinpointing direct responsibility for decisions made by autonomous machines. These failures were not fleeting laboratory experiments; they involved malfunctioning software agents, the exposure of sensitive data, outright errors in technology-assisted decisions, and security breaches that prompted nearly half of the affected companies to rewrite existing governance policies.

A study conducted by Sensex Wide for IBM, covering 3,221 senior executives in organizations with at least 1,000 employees, found that 75 % of the UAE’s largest companies experienced AI-related incidents that formally reached senior leadership. In response to those operational shocks, 49 % of executives adjusted governance policies and frameworks, while 40 % altered investment pathways and application rollout plans. Yet the shocks did not halt the push toward automation; 60 % of corporate officers affirmed their continued expansion within agreed-upon risk limits while building supervisory frameworks, whereas only 30 % preferred to postpone expansion until the governance system was complete.

The rapid move to full production exposed a deep gap in accountability and remediation rules.Software agents are no longer just chat models responding to employee commands; they have become systems with degrees of autonomy to execute complex procedural tasks. According to the data, AI agents have entered full production in at least one role at 38 % of the surveyed companies and have been widely embedded in operational processes at 13 % of them. With 78 % of executives expecting agents to have self-directed decision-making autonomy by 2027, the lack of consensus on who is accountable for their errors becomes a deferred administrative and legal minefield.

When the critical question of who bears the blame was raised, leaders in the UAE gave fragmented answers: 21 % saw the chief risk or compliance officer as the primary accountable party, 17 % viewed responsibility as shared among certain senior executives, and 15 % placed the blame on the head of technology or information. Legally, 26 % believe that assigning liability for severe damages depends on the circumstances of each case, 19 % stated that the current regulatory arsenal is entirely insufficient to address these disputes, and 16 % called for liability to be split between the user company and the technical vendor that developed the system.

This contractual ambiguity now demands a fundamental shift in how tech deals are drafted across the region.Technology directors and legal officers in Gulf and Egyptian markets find themselves urgently needing to redesign service-level agreements with foreign tech giants, as reliance on an external agent for credit decisions or customer data sorting means the local firm alone bears legal costs and reputation risk if the agent fails. Thirty-seven percent of corporate officers call for new legal and contractual frameworks that clearly define supplier obligations, while 36 % advocate for cross-jurisdictional governance mechanisms, and the same share insists on possessing an internal safety key that enables immediate and complete freezing of all agent activities within the organization at the first sign of deviation.

The study also shows that converting these risks into precise financial figures remains at a nascent stage; only 39 % of organizations have conducted an official quantitative assessment of the financial risks stemming from AI governance gaps for presentation to boards. Conversely, government regulations remain the strongest driver of protective spending; 59 % of executives attribute their supervisory investments to the UAE’s national AI strategy, ahead of international regulatory compliance at 42 % and sector-specific regulations at 38 %.