UK CIOs face AI blame as sovereignty drives buying

Research published by 8x8 shows that half of UK technology leaders say the Chief Information Officer is personally accountable when an AI agent makes a mistake. The figure rises to 62% at organisations with more than 500 employees.

The findings are based on a survey of 2,501 Chief Information Officers and Chief Technology Officers in the UK, US, France, Australia and the Republic of Ireland, conducted by Censuswide for 8x8.

The UK results point to a widening gap between responsibility for AI decisions and the systems used to manage risk. Respondents most often said the Chief Information Officer is held accountable when an AI agent fails, while far fewer pointed to legal or compliance teams.

In the UK, 50% of respondents said the Chief Information Officer carries personal accountability when an AI agent makes an error, compared with 20% for customer service leadership and 5% for legal or compliance. Among government and public administration respondents, 86% placed that responsibility on the Chief Information Officer, the highest level reported by any sector in the survey.

The pattern also shifts with scale. At organisations with 50 to 99 employees, 45% of respondents said the Chief Information Officer is personally accountable, with the share rising sharply among larger employers.

Jamie Snaddon, EMEA managing director at 8x8, said the UK figures suggest the effects of AI mistakes are extending beyond IT departments.

"What jumps out in the UK data is how many hands are already on this," said Jamie Snaddon, EMEA managing director at 8x8. "Half of the CIOs we spoke to say they personally carry the can when an AI agent gets it wrong, but one in five also name customer service leadership. That's higher than anywhere else we surveyed. That's not confusion. The fallout from AI is being felt more widely across the business than most people assume."

Data concerns

The survey also found that UK organisations place unusually high weight on data sovereignty when choosing AI suppliers. Some 39% of UK Chief Information Officers said the location of AI infrastructure is now the primary criterion in vendor selection, a larger share than in any other market covered by the study.

The concern was even more pronounced at larger businesses, where 62% of respondents at companies with more than 500 employees said sovereignty was a primary factor in vendor choice. Separately, 91% of UK technology leaders said data sovereignty had risen up their list of priorities over the previous 12 months.

The issue appears tied to a broader rise in regulatory scrutiny and concern over where data is processed and stored. In that context, vendor selection is being shaped not only by price or functionality, but also by where AI systems and associated data are located.

"UK CIOs are telling us that where their data lives now decides who gets their business," said Snaddon. "Sovereignty is the opening question in almost every conversation we're having with customers here."

Consolidation barriers

The research found little appetite among global technology leaders to argue that one platform could not, in principle, meet their needs. Just 9% of Chief Information Officers worldwide said no single platform can do so, suggesting the main barriers to consolidation lie elsewhere.

Migration costs and complexity, regulatory and data residency constraints, and internal politics were cited more often than technical shortcomings. In the UK, respondents also highlighted vendor lock-in as a distinct obstacle, particularly in the public sector.

Among government and public administration respondents in the UK, 71% said vendor lock-in from existing contracts was the biggest obstacle to consolidation. That was more than three times the UK-wide figure, indicating that older procurement arrangements may be limiting efforts to simplify communications and AI systems.

For many technology leaders, the challenge is no longer simply whether AI tools can be deployed. It is whether existing contracts, operating models and governance structures allow organisations to control how those tools behave when something goes wrong.

"The lock-in figure doesn't surprise me," said Snaddon. "A lot of UK organisations, especially in the public sector, are stuck in contracts signed before AI was a board-level conversation. The technology to consolidate already exists. What's missing is a commercial path out of those agreements that doesn't put all the migration risk on the customer."

The findings suggest accountability for AI in UK organisations is concentrating around the Chief Information Officer role at the same time as vendor choice, governance and compliance demands are becoming harder to separate. In larger organisations, and in the public sector in particular, that combination is making communications infrastructure a central point of control as well as a source of operational risk.