Research released today by ICS.AI suggests UK employees are broadly supportive of artificial intelligence in the workplace, but remain firm that responsibility for decisions must sit with humans.
The nationwide survey of 2,000 employees found that 62% are excited about the prospect of AI at work. However, more than eight in ten employees expect human involvement when AI is used to support important workplace decisions.
The largest group (46%) believe AI-generated recommendations should be signed off by a person, while a further 37% expect humans to make final decisions supported by AI. Only 7% say they are comfortable with fully autonomous AI decision-making.
The findings point to a workforce that is open to automation, but cautious about delegating accountability to systems.
Adoption levels across organisations remain uneven. One in five organisations are not using AI at all, while around a third (34%) describe themselves as “AI-curious”, running small pilots or isolated experiments. Just over a third (36%) have partially embedded AI across some teams.
Only one in ten organisations can be described as “AI-native”, with AI deployed in a joined-up, organisation-wide way.
That fragmentation may explain mixed perceptions of value. Nearly three in ten organisations (29%) report already seeing returns from AI investment, while the largest group (40%) say it is too early to tell. Employees also indicate that benefits are either unevenly distributed or not yet fully realised.
Current usage is largely operational rather than strategic. The most common applications are administrative tasks such as document drafting and data entry (44%). Fewer than one in three employees say AI is being used for planning, forecasting, or insight-driven decision-making.
Comfort levels also vary. Two thirds (67%) say they feel at least somewhat comfortable using AI at work. Yet one in three say they do not feel prepared at all.
Martin Neale, CEO of ICS.AI, said: “What this research really shows is that point-to-point AI use will never deliver true transformation. While many employees are comfortable using AI in parts of their work, a full third don’t yet feel prepared at all – and that gap becomes critical as AI takes on more influential roles.
“Human-in-the-loop models are essential at this stage. They allow organisations to move beyond isolated pilots and scale AI responsibly, with clear accountability, shared understanding and trust. The findings suggest that the next phase of the UK’s AI journey will be defined not by how quickly organisations adopt new tools, but by how effectively they align technology, people and accountability to build trust at scale.”
The research suggests the next stage of AI adoption in UK workplaces will hinge less on experimentation and more on governance, skills development, and clearly defined lines of responsibility as deployment expands into higher-impact roles.
