UK tech workers demand better AI training from employers

UK tech workers are calling for better AI training from employers, according to research by La Fosse. The study found a gap between widespread daily use of AI and the level of formal instruction staff receive.

A survey of 2,000 workers found that 92% of UK tech professionals use AI every day, while 87% said their employer offers some form of AI guidance. Yet only 58% had received formal training, and many said it was too basic for their roles.

The findings suggest a shift in what employees want from AI learning. Just 23% said they need training in AI fundamentals, while more want practical instruction tied to the tasks and risks of their specific roles.

Cybersecurity and data privacy ranked as the top area for further training, cited by 39% of respondents. Data analysis and visualisation, and data quality and integrity, followed at 34%.

As AI becomes routine across workplaces, the mismatch is creating operational and governance concerns. Only 37% of tech workers said they always verify AI outputs before using them, while 67% had seen AI cause a mistake in their organisation.

Among senior leaders, 29% of C-suite respondents said AI errors had caused serious business impact. The study also found uneven access to training across organisations, with entry-level and intermediate staff less likely to receive formal support than senior executives.

Entry-level employees were the least likely to have received formal training, with 26% reporting none at all. Among intermediate workers, 34% said they had no training, compared with 6% of C-suite leaders.

That matters because junior and mid-level staff are often among the most frequent users of AI tools in day-to-day work. In many businesses, they use AI to rewrite emails, summarise notes and support routine decisions, even as employers have yet to build structured training around those tasks.

Claudia Cohen, director of La Fosse Academy, said the pace of workplace change is outstripping internal training efforts.

"What we're seeing is that AI is no longer sitting alongside roles, it's actively reshaping them. The expectation of what 'good' looks like in a role is changing faster than most training programmes can keep up," she said.

Cohen said many employers have adopted AI tools without building the skills needed to use them properly.

"Right now, most organisations have AI usage without AI capability. People are using the tools, but not in a way that fundamentally improves how work gets done and that's where both the risk and the missed opportunity sit," she said.

Role-specific use

The research suggests workers increasingly want training linked to their actual functions rather than broad introductions to AI. Cohen pointed to examples across legal, learning and development, and operations teams where AI could be used in more targeted ways.

"For someone in legal, it could mean reviewing and stress-testing policies at scale. For someone in learning and development, it could mean designing and tailoring training programmes in a fraction of the time. For someone in operations, it could mean improving decision-making through better use of data. The gap is understanding how to apply AI effectively within a role," she said.

She argued that employers need to move away from one-off training sessions and towards continuous learning that reflects the realities of each role. That includes teaching staff how to apply AI to specific tasks, how to interpret outputs and where the risks lie in their workflows.

"Organisations need to move beyond generic AI training and focus on real-world application. In the tech industry overall, the most in-demand AI skills are cybersecurity, data privacy, data analysis and data integrity. These more specialised areas show the need for a deeper, practical training approach tailored to departments and even individual roles.

"While foundational knowledge has its place, helping employees understand how AI fits into their specific roles is hugely important. Businesses need to train staff on how to apply AI to everyday tasks, how to interpret outputs and where the risks lie in their workflows. Without that context, even the most advanced tools can be misused or underused. The value of AI isn't in understanding it conceptually, it's in how it's applied. Without that translation into day-to-day work, adoption remains surface-level."

Controls and training

Cohen also said businesses need stronger checks on AI-generated material as use widens across organisations.

"As AI adoption accelerates, verification can't be treated as optional. Our research shows a growing tendency to trust outputs without question, which is where business risk begins to grow. Organisations need to embed clear rules into everyday processes, whether that means mandatory review steps, defined accountability or stronger data governance policies.

"There's a growing tendency among employees to trust outputs too quickly. But hallucinations, bias and data quality issues in AI are still very real. If employees don't understand how their inputs affect outputs, and how those outputs are then used across the business, risk compounds quickly."

She said the people using AI most often should be the priority for formal instruction.

"There's a clear imbalance in who receives AI training within an organisation, and it's creating unnecessary exposure. Entry-level and mid-level employees are often the most frequent users of AI tools, yet they're the least likely to have received formal support. These are the individuals making day-to-day decisions with AI, so equipping them with practical skills will have a greater impact. Alongside reducing risk, targeted investment unlocks value from AI across the organisation.

"Organisations need consistent, embedded learning, practical sessions, team discussions and time to apply lessons in real scenarios. Importantly, they also need to create an environment where people feel comfortable experimenting.

"Right now, we see three distinct behaviours: people who are leaning in and experimenting with AI, people who are letting it happen around them, and people who are actively avoiding it. The risk is that those in the latter two groups quickly fall behind as roles evolve."

On the broader response required from employers, she added: "To tackle this education gap, employers need training that reflects real tasks, real decisions and real risks within their organisation. This should become part of how people work, not something they dip in and out of, and it must be specific to each role within the business.

"But that doesn't remove the need for fundamentals - it actually makes them more important. AI can accelerate experience, but it can't replace judgement. Without that balance, organisations risk moving faster, but in the wrong direction."