New research has revealed that nearly half of UK Chief Information Officers (CIOs) feel personally responsible when an AI agent makes a mistake, even if they are more of a symbolic leader than an active decision-maker in their department. The study, conducted by the company 8x8, found that the size of a company can influence the level of personal accountability placed on a CIO. Larger companies, with more employees and managers, often expect their CIOs to take on greater personal responsibility for AI-related errors. Among businesses with between 50 and 99 employees, 45% of CIOs reported feeling personally accountable for AI mistakes. However, this number rose to 62% in companies with 500 or more employees. Government and public administration organizations placed the highest pressure on CIOs, with 86% of them feeling personally responsible for AI errors. This suggests that regulatory and public scrutiny may be driving a higher level of accountability in these sectors. CIOs are also facing challenges related to data sovereignty, which refers to the control a country or organization has over its data. Nearly 40% of CIOs said the location of AI infrastructure is now the most important factor in choosing which platforms to use. Concerns about where data is stored and processed are now central to the decision-making process, both at the start of procurement and in broader strategic discussions. Almost all (91%) UK technology leaders believe that digital sovereignty — or the control over digital assets and data — has become a more important priority this year. Other factors influencing their decisions include vendor lock-in, cost, and the complexity of managing AI systems. Jamie Snaddon, Managing Director for EMEA at 8x8, noted that questions about sovereignty are often the first topic raised in conversations with customers. With the increasing pressure from accountability, data concerns, and regulatory expectations, the role of a CIO is becoming more complex and risky. The research suggests that CIOs may be more hesitant to adopt AI technologies and might even be less inclined to take on the position in the future, unless the current challenges are addressed.