Most organisations assume that AI risk sits with the people furthest from the decision. Junior staff experimenting. Frontline workers cutting corners. But new independent research with over 2,000 UK tech workers tells a different story. 

The biggest AI risk-takers in UK businesses are senior leaders. The people with the most influence, the most autonomy, and the highest business impact are also the ones most likely to push the boundaries of responsible AI use. And the consequences, when things go wrong, are proportionately larger. 

The pattern is consistent across the data

The research, conducted by Censuswide on behalf of La Fosse, found that C-suite executives are not just heavy AI users. They are using AI in ways that create material risk for their organisations. 

Uploading confidential data into AI tools: 

  • C-suite: 73% 
  • Directors: 74% 
  • Intermediate staff: 35% 
  • Entry-level staff: 42% 

Senior leaders are roughly twice as likely to put sensitive information into AI systems as those further down the organisation. And in most cases, this is not accidental. Of the C-suite executives who uploaded confidential data, 53% did so knowingly. 

73% of C-suite executives have uploaded confidential company data into AI tools 

Using AI beyond their area of training 

The confidence that comes with seniority does not always translate into AI competence. 

  • 78% of C-suite executives have used AI to complete work they were not trained to do 
  • 73% of directors have done the same 
  • Among intermediate staff, that figure drops to 42% 

Senior leaders are applying AI to complex, consequential tasks while operating outside their area of training. The research confirms the business impact: 40% of C-suite executives report having experienced a serious negative outcome from AI use, more than any other group in the study. 

78% of C-suite executives have used AI to complete work they were not trained to do 

The governance gap starts at the top

Governance frameworks are designed to prevent exactly this kind of risk-taking. But they only work if they apply to everyone. When senior leaders routinely operate outside them, it signals to the rest of the organisation that the rules are optional. 

The research found: 

  • 37% of tech workers say AI governance ownership is unclear in their organisation 
  • 7% say nobody is responsible at all 
  • 66% of C-suite executives know who is responsible for AI governance 
  • That falls to just 30% of intermediate employees 

When the people setting the rules are also the ones most likely to break them, governance stops functioning as a control and becomes a formality. 

Why this matters beyond the boardroom 

The risk is not contained to individuals. When senior leaders make AI-assisted decisions based on inaccurate outputs, or operate outside approved tools, the downstream effects are felt across teams, clients, and sometimes regulators. 

As AI influences regulated decisions, financial disclosures, and strategic direction, accountability follows the person who relied on the tool, not the tool itself. Boards and senior leaders need to be asking hard questions about their own AI behaviour, not just that of their teams. 

What good AI governance actually looks like

Closing this risk requires more than a policy update. It requires: 

  • Governance embedded in how work gets done, with clear accountability at every level including board level 
  • Honest assessment of where senior leaders are using AI beyond their competence 
  • A culture where responsible AI use is demonstrated from the top, not just required at the bottom 
  • A named AI leader with board-level authority and real technical depth 

Inovus works with organisations to design AI governance models that align leadership accountability with operational reality. That means governance built into day-to-day decision-making, not written into documents that sit unread on an intranet. 

Take the next step

If you want to assess where your AI governance has gaps, particularly at senior level, we offer a free 30-minute consultation to discuss your current approach. 
Book your free consultation

Read the full research

This article draws on findings from AI in the Workforce: The Hidden Risk for UK Businesses, independent research with over 2,000 UK tech workers. 
Download the whitepaper