PwC survey finds business and tech leaders split over who owns AI risk at work
PwC research shows executives disagree on who should take responsibility for AI risk, with the question of a dedicated AI chief still unresolved.
Business and technology leaders cannot agree on who should own AI risk inside their organizations, according to new PwC research that points to a widening accountability gap as AI tools spread through workplaces. The survey, reported by ZDNet, found the two groups diverging on where responsibility should sit, leaving companies without a clear answer at a moment when regulators and boards are asking harder questions about AI failures.
The split matters because AI systems now touch hiring, credit, customer service and internal decision-making, and each of those uses carries a different kind of exposure. A model that misroutes a support ticket is a nuisance; one that screens out job applicants or denies a loan is a legal and reputational problem. When business and technology teams each assume the other is watching, nobody is, and the gap tends to surface only after something goes wrong.
PwC's findings echo a broader pattern across the industry. Companies have rushed to appoint chief AI officers, but the title varies widely in scope. Some hold budget and strategy; others function as evangelists with little authority over engineering, legal or compliance. The survey suggests that ambiguity is not a cosmetic problem, since accountability without authority rarely produces real oversight.
The question of whether a single AI chief is the answer remains open. A dedicated executive can consolidate reporting lines and give boards one throat to choke, but critics argue that concentrating AI risk in one seat lets every other function off the hook. Data protection officers, chief information security officers and general counsel already own adjacent risks, and AI failures often cut across all of them at once.
What the research does make clear is that the current arrangement is unstable. As AI regulation advances in the European Union and several U.S. states, and as insurers begin asking pointed questions about model governance, companies will need to name an owner rather than assume consensus. The next signal to watch is whether boards start writing that responsibility into formal charters, or whether the disagreement PwC found simply persists until an incident forces the issue.