New survey of 2,400 execs and employees within large enterprises suggests up to 5,557 of the UK’s largest businesses – representing 27% of the UK’s workforce – may be at risk as speed of adoption outruns safeguards
As the race to adopt AI continues to accelerate, major new research from WRITER, the leading AI agent platform for the enterprise, reveals a compounding set of AI data breach, security, ethics and governance failures driven by the intensity of AI adoption pressure. The findings highlight the urgent need for stronger AI leadership across organisations to make the technology’s rollout a success while preserving employee trust and operational compliance.
One in three employees (35%) admit to having fed proprietary, confidential or sensitive company information into a public AI tool, a behaviour driving what many executives suspect is a wave of silent AI data breach. In fact, two-thirds of C-suite leaders (67%) believe their organisation has already suffered a data leak or security breach caused by an employee using an unapproved AI tool. Only a third of executives are certain that no breach has taken place. The findings suggest that as many as 5,557 large UK enterprises may have already exposed themselves to risk, potentially making employees and customers vulnerable.
Also Read: How CSOs Can Bridge The Skills Gap: Balancing Rapid AI Innovation with Robust Cybersecurity
Research suggests that this concern is well founded, given the prevalence of “shadow AI” within many businesses. 16% of employees admit to using AI tools explicitly banned by their employer. The top reason cited by employees using such tools are a willingness to “use whatever it takes to get their work done” (40%), followed by company-approved tools being too poor to use (32%), as well as a lack of enforcement on which tools get used. At the same time, more than a third of executives (35%) concede that they do not have full visibility or control over which AI tools their employees are really using.
A Culture of Fear Around Harmful AI
More than a quarter of employees (28%) say they have witnessed an AI tool at work produce a result that was dangerously wrong, unethical, or biased. Yet a significant proportion are reluctant to flag it, with three out of ten employees feel unsafe reporting dangerous or unethical AI behaviour to their employer, fearing retaliation. However, there appears to be a disconnect between ground-level reality and management perception, as the majority of executives (90%) believe employees are safe to speak up.
The findings raise serious questions about whistleblower protections and internal reporting cultures at a moment when AI tools are getting more deeply embedded in mission-critical workflows. In fact, the pressure placed on employees to use AI is itself a driver of this unwillingness to flag harmful AI as staff fear being seen as obstructing AI adoption if they raise concerns.
Also Read: Redefining Insurance with the Power of AI and Human Judgement
Agentic AI Governance has High-Stakes but Gaps Remain
Autonomous AI agents are moving from experiments to operational reality faster than most companies’ guardrails can keep up. This is reflected in the C-suite’s inability to deactivate agents, as more than a third of executives (35%) say they are not confident they could “pull the plug” if an agent went off the rails and started causing financial or reputational damage. This is despite the fact that the perceived consequences for such an agent-driven error are high for executives. Six out of ten leaders believe this would cost a senior executive their job, most commonly citing the CEO (30%) or the CIO/CTO (29%) as most likely to be impacted. In addition, 36% admit their organisation still lacks a formal, documented plan for supervising agents at all.
Asked to identify the biggest governance challenges, leaders most frequently cited security and data protection (45%), followed by employee training (35%), transparency about how agents operate (31%), and explainability (30%). And despite widespread evidence of problematic AI outputs, only 25% of executives ranked ethical alignment, ensuring agents behave ethically, as a top governance concern. Finally, with three quarters of 75% of executives saying their company’s AI strategy is driven more by public signalling than by practical internal direction, this reinforces a lack of confidence that governance commitments are being taken seriously.