While many organisations are still grappling with how to best integrate AI into their workflows, a quieter revolution is taking place within the workforce. Employees are increasingly using tools like ChatGPT, generative AI and other automation platforms with little or no oversight from management. This covert use of AI is inevitable. Handled badly, it’s a double-edged sword.
The AI hype vs reality
With all the buzz around AI, from leaders declaring “We need AI!” (to do what, they’re not sure) to the widespread adoption of tools like Copilot and ChatGPT, it’s easy to assume that everyone has a deep understanding of what AI can and can’t do. But just because AI is a hot topic doesn’t mean it’s fully understood or being used properly. Adoption is moving faster than most organisations can grasp the risks and responsibilities that come with it, or set up the governance and support it needs. It’s not just about having the technology in place; it’s about creating a framework that ensures employees understand how to use it safely and responsibly. Remember: IT is never just tech.
Covert AI adoption
Like it or not, employees are turning to AI tools to boost productivity, streamline tasks or assist with decision-making. Much like when tech-savvy employees create unofficial business-critical systems outside IT’s control (and, worryingly, sometimes with its knowledge), AI tools are being adopted without formal oversight. Without a full understanding of how they work, they’re using AI without considering long-term implications for data security, compliance and decision-making quality.
Even when organisations provide sanctioned AI tools, employees may default to external ones due to familiarity or convenience. This creates a ‘shadow AI’ problem, where company-approved tools exist, but unapproved alternatives remain in use. To counteract this, businesses must ensure their AI offerings are accessible, user-friendly and clearly more beneficial in terms of security, compliance and integration.
Potential pitfalls for businesses
- Data security and privacy risks: AI tools often handle sensitive data. Using them without a clear data governance framework can lead to data breaches. Employees may unknowingly expose company information to unvetted platforms or inadvertently violate compliance regulations.
- Misinformation and accuracy concerns: AI, especially generative models like ChatGPT, can produce information that seems convincing but may not always be accurate. Inaccurate information can lead to flawed decision-making, potentially causing reputational damage, financial loss or operational disasters.
- Bias and ethical issues: AI tools can introduce biases or reinforce existing ones. Without understanding the underlying algorithms, employees may unknowingly amplify biases in decision-making or inadvertently contribute to unfair practices.
The need for AI awareness and governance
Acknowledge. Incorporate. Integrate.
- Acknowledge it’s already there: Companies need to face the reality that employees are already using AI tools and recognise that this trend isn’t going away. The first step is acknowledging the need to establish guidelines and frameworks for AI usage in the workplace. Banning it isn’t the answer.
- Incorporate AI in data strategies: AI should be embedded into broader information management and data strategies. This includes ensuring that AI usage aligns with data security protocols, ethical standards and compliance regulations.
- Integrate AI into workplace policies: AI governance belongs in SyOps and technical policies. It also belongs in employee handbooks, training and workflows. Employees need clear guidance on responsible AI use, data handling and oversight. Making AI usage policies accessible and explicit keeps its use safe, ethical and aligned with business objectives.
Who’s at fault when it goes wrong?
Beyond governance, organisations must clarify accountability. If an AI-generated decision leads to a mistake, who’s responsible: the tool, the employee or the company? Without clear accountability, businesses risk legal and ethical grey areas. AI policies must define ownership, establish human oversight and ensure that responsibility remains with decision-makers, not algorithms.
Educate!
AI governance, like all governance, isn’t a one-time effort; it requires ongoing adaptation. As AI tools evolve from simple chatbots to autonomous decision-making agents, companies must continuously reassess their policies, training and risk management strategies. The AI landscape moves fast; an organisation’s ability to stay ahead of these changes will define whether it remains in control or simply reacts to AI-driven disruptions.
- Training on data security: Staff should be educated about the risks associated with AI, such as data leakage, security vulnerabilities and how AI tools might interact with sensitive company data.
- Promote awareness of AI limitations: It’s crucial for employees to understand the limitations of AI, such as potential inaccuracy and biases, so they don’t rely solely on these tools without cross-checking the information.
- Encourage responsible AI use: Provide employees with clear guidelines on when it’s appropriate to use AI tools and when manual review or human intervention is necessary.
Conclusion
AI in the workplace is a growing reality. Companies can’t afford to ignore it. I know it’s boring, but by adopting AI responsibly, putting governance in place and educating employees, organisations can reduce the risks and still get the full value of AI. The first step is acknowledging the covert use of AI already happening within your organisation and developing a proactive strategy to ensure it’s used ethically, securely and effectively.