UK companies are increasing board-level oversight and tightening security controls around artificial intelligence as the technology moves from experimental projects into core business operations, according to research from KPMG.

Almost two-thirds of UK organisations surveyed, or 64 per cent, said accountability for decisions informed by AI now sat at C-suite level or higher, compared with 54 per cent globally.

The findings form part of KPMG’s latest Global AI Pulse, which surveyed 2,131 senior leaders across 20 countries between late July and August, including 100 in the UK. The research suggests companies are increasingly focusing on how AI systems are governed and measured rather than simply whether the technology can be deployed successfully.

Globally, 55 per cent of organisations now operate formal infrastructure for managing AI, according to KPMG. That rises from 31 per cent among companies still experimenting with the technology to 86 per cent among those reporting established returns on their investments.

The shift is also being driven by security concerns. Eight in 10 UK respondents said they were adapting their cyber security operations and strengthening policies and controls as AI adoption increased.

The issue has attracted growing attention from policymakers. The UK government's voluntary AI Cyber Security Code of Practice sets out 13 principles covering areas including human responsibility, infrastructure, supply chains, testing and monitoring. It is intended to form the basis for an international standard.

Companies are also paying greater attention to where their models and data are hosted. KPMG found 69 per cent of UK respondents considered “model sovereignty” when making AI decisions, with 24 per cent operating an organisation-wide strategy that was regularly reviewed.

However, financial oversight appears less developed. While 67 per cent of organisations review AI costs when approving projects and 58 per cent monitor them during operation, just 13 per cent consistently assess value against cost across their organisation.

Paul Henninger, head of technology and data at KPMG UK, said deploying AI was “only the first step”, with companies increasingly needing leadership and governance structures capable of supporting the technology at scale.

Employee use of autonomous or semi-autonomous AI agents is meanwhile continuing to expand. Some 35 per cent of UK organisations reported significant adoption, up from 30 per cent in the first quarter.

KPMG said the findings indicated that competitive advantage was increasingly shifting from access to AI towards companies' ability to manage the technology effectively.