Recently, it came to my attention that a member of our team had been using AI to complete some tasks.
She is hardworking, and I had no reason to doubt that her intentions were good. Still, the discovery made me pause and ask an important question: should staff be allowed to use AI to do the work they were hired to perform?
Does using AI show initiative and help a good employee work more efficiently, or does it create concerns about quality, confidentiality and accountability?
This is no longer a question business owners can ignore. Staff are already using AI to draft emails, summarise documents, prepare proposals, generate ideas and solve everyday problems. The answer is not to panic or pretend it is not happening. The business needs to decide how AI should be used.
AI Can Make Good Staff More Productive
Used properly, AI can reduce repetitive work and give employees more time for tasks that require judgment, creativity and customer understanding.
A customer-service employee can use it to prepare a first draft of a response. An administrator can turn meeting notes into action points. A marketer can generate alternative headlines before choosing and improving one. A researcher can use it to organise questions for further investigation.
The employee must still understand the work. AI should help them think and execute faster, not produce something they cannot explain or defend.
The Risks Are Real
AI can produce confident answers that are incomplete, misleading or completely wrong. If staff submit the output without checking it, the business remains responsible for the mistake.
There is also a confidentiality problem. An employee may paste a customer’s personal information, an unpublished proposal, staff records, financial figures or a contract into a public AI tool without understanding where that information may be stored or how the service handles it. Nigeria’s data-protection guidance expects businesses using AI to protect personal data and assess the risks involved.
Other risks include unoriginal content, inconsistent communication, overdependence and the gradual loss of staff skills. A junior accountant who accepts an AI-generated calculation without understanding it has not saved time. The error has merely been hidden.
Do Not Choose Between a Ban and Unlimited Use
A complete ban may push AI use underground. Staff may continue using personal accounts and devices, leaving the business with even less visibility. Unlimited use is equally risky.
A better approach is to divide work into three categories.
Permitted: Staff may use approved AI tools for brainstorming, outlines, first drafts, formatting and summarising non-confidential information. They must review and take responsibility for the result.
Approval required: A supervisor must approve AI use for customer-facing advice, important proposals, financial analysis, recruitment decisions, contracts and compliance-related work. These tasks need stronger checking because an error can affect money, people or legal obligations.
Prohibited: Staff must not enter passwords, customer records, bank details, private contracts, confidential business plans or sensitive employee information into an unapproved AI tool. AI should not make final hiring, disciplinary, lending or other high-impact decisions.
Create a Simple AI Policy
Your policy does not need to be twenty pages. Start with one page answering five questions:
- Which AI tools may staff use?
- Which tasks are permitted, restricted or prohibited?
- What information must never be uploaded?
- Who must review AI-assisted work?
- What should an employee do if sensitive information is entered accidentally?
Provide a business-managed account where possible, train staff with real examples and review the policy as the tools change. Most importantly, make it clear that using AI does not transfer accountability. The employee who submits the work must check the facts, calculations, tone and instructions.
Final Word
Discovering that a good employee used AI is not automatically evidence of laziness or dishonesty. It may show initiative. It may also expose a gap in the way your business manages technology.
Do not ask only, “Did you use AI?”
Ask: “Was it an approved tool, was the information safe, was the output checked, and can you stand behind the final work?”
Those questions allow your staff to use technology without allowing technology to create unnecessary risk for the business.
