Kenya has no law written specifically for AI in hiring. That does not make the work risk-free. Every CV an AI tool reads is personal data, every shortlist is a decision about a person, and the courts already have firm views on unfair process. HR teams are the ones who answer for it.
Here are five checks to run before you automate anything.
1.Know what the tool does with candidate data
CVs, ID numbers and phone numbers are personal data under the Data Protection Act 2019. The Office of the Data Protection Commissioner (ODPC) has issued an AI guidance note, and it has ruled on employee data complaints before. Before you upload a batch of CVs to any tool, find out where the data goes, who can see it, and how long it is kept. If you cannot answer those three questions, do not upload.
2. Keep a person in charge of every outcome
Let software do the repeat work, such as sorting, matching and flagging. A named person should make the decision to reject or advance a candidate, and should be able to say why. This is the simplest protection you have.
3. Test for fair outcomes
A shortlist that always favours one university, one county or one gender is a problem even if nobody meant it. Run a sample of past applications through the tool and look at who comes out. Write down what you found and what you changed.
4. Get statutory deductions right before you automate payroll
Payroll is where small errors become costly. From 1 February 2026, NSSF Year 4 applies: 6% from the employee and 6% from the employer, on pensionable pay up to KES 108,000, so the maximum is KES 6,480 each. SHIF is 2.75% of gross pay with a minimum of KES 300, and the Housing Levy is 1.5% each side. SHA charges 2% a month on late SHIF. Check your system’s rates against the latest notices from NSSF, KRA and SHA, not against last year’s setup.
5. Treat redundancy files as evidence
On 25 June 2026 the Employment and Labour Relations Court found that Nokia Solutions and Networks had declared a senior manager redundant unfairly and awarded KES 9.8 million. The employer had to show that the job really became unnecessary and that it ran a fair selection among similar roles. If an AI tool helped pick who goes, your records need to show the criteria and the human decision behind them.
What about the global picture?
The EU’s AI Regulation (2026/1744) classes tools that filter applications or evaluate candidates as high-risk. It is written for the EU, so it is not Kenyan law, but it shows where regulation is heading. SHRM’s 2026 survey of 1,908 HR professionals found that 46% of organisations expect to use AI in HR this year, and recruiters lead at 27%. Many Kenyan teams will be using these tools soon, so the habits are worth building now.
Where Vunoh Global fits
We help HR teams set up AI for repeat tasks such as screening, onboarding queries and policy questions, with the checks above built in. If you want to look at one process first, get in touch at info@vunohglobal.com.