Industry8 min read

How Can Lab Technicians Use AI Safely?

On this page

Lab technicians can use AI safely for work that holds no patient information: drafting SOPs, turning approved specimen instructions into plain-word leaflets, sorting quality control and stock sheets, and revising for CPD. AI should never interpret a result, validate or release a report, or suggest a diagnosis. Those calls stay with qualified laboratory staff and the clinician, because AI can sound certain and still be wrong.

This guide is for medical laboratory technicians, technologists and lab managers in Kenya, in hospital, private and research laboratories. It is general information, not clinical, regulatory or legal advice.

Who regulates lab work in Kenya

The regulator is the Kenya Medical Laboratory Technicians and Technologists Board (KMLTTB), set up under the Medical Laboratory Technicians and Technologists Act, 1999 (Cap. 253A). Section 5 of the Act gives the Board "general supervision and control over the training, business, practice and employment" of laboratory technicians and technologists. It also licenses practice and regulates professional conduct.

Three things from the Board's own documents matter here:

  • Private practice needs a practising certificate and an annual licence under sections 20 and 21 of the Act. The practising year runs from 1 January to 31 December (section 23).
  • CPD is tied to your licence. The KMLTTB CPD guideline (revision dated March 2024) says practitioners need at least 50 CPD points each calendar year. You need those points to renew your licence.
  • No AI-specific rule appeared on the KMLTTB pages we checked on 15 November 2026. The CPD guideline does list AI in medical laboratory practice as one CPD topic. Your existing duties still apply to any tool you use.

The data rules are just as important. Kenya's Data Protection Act, 2019 defines health data to include data collected while registering for or receiving health services. Health status and genetic data are both sensitive personal data. The data commissioner's Guidance Note on the Processing of Health Data (December 2023) names laboratories and Laboratory Information Systems directly.

The World Health Organization's 2024 guidance on large AI models warns of "false, inaccurate, biased, or incomplete statements". It also warns of "automation bias", where people miss errors because they trust the tool. In a lab, that is the risk to design around.

Six ways lab technicians can use AI today

Each prompt uses the four Cs we teach at FuKazee: Clarity, Context, Constraints and Customisation. None of them needs a patient's details.

1. First drafts of SOPs

AI is good at structure. It does not know your analyser, your kit insert or your accreditation scope, so it must not invent steps.

Clarity: Draft a standard operating procedure for [task, for example receiving and logging specimens at the reception bench]. Context: a hospital laboratory in Kenya. Constraints: headings are purpose, scope, responsibilities, materials, procedure, records, references. Use only the steps in my notes. Where a step depends on a kit insert, analyser manual or national guideline, write [CHECK SOURCE] instead of guessing. Under 700 words. Customisation: numbered steps short enough to print and pin at the bench. MY NOTES: [paste your own rough steps]

Check: the quality manager compares every step with the manufacturer's instructions and your current documents, then signs it off through your normal document control.

2. Patient preparation leaflets in plain words

Patients often arrive unprepared for fasting tests or timed samples. AI can turn approved instructions into something short.

Rewrite the approved patient instructions below as a half-page leaflet for adults in Kenya. Use only these facts. Do not change any time, number or warning. Add nothing new. Mark anything unclear as [LAB TO CHECK]. Then give the same points in natural Kiswahili. TEXT: [paste your laboratory's approved instructions]

Check: read every line against the source, especially times and fasting rules. A fluent Kiswahili speaker checks the translation.

3. Quality control and turnaround sheets

Internal QC results and instrument logs normally hold no patient data, which makes them a safe place to start. Turnaround data is safe only once sample numbers and names are removed.

Below are last month's internal QC results for [analyte] on [analyser], with columns: date, level, result, target mean and SD. Show which days fall outside 2 SD and 3 SD, and any run of results on one side of the mean. Present it as a table. Then explain how to build the same check in Excel. Do not tell me whether to accept or reject any run. DATA: [paste]

Check: recalculate a few points yourself. Accepting or rejecting a run follows your lab's QC rules and the supervisor's judgement. For spreadsheet help, see how to use Copilot in Excel.

4. Reagent stock and expiry

Here is our reagent and consumables list with columns: item, lot, quantity, average weekly use, expiry date. List items likely to run out within three weeks, lots expiring within 60 days, and anything we hold more than four months of. Three tables, then the Excel formulas to repeat this monthly. LIST: [paste, with supplier prices removed if confidential]

Check: count a few items on the shelf and test the formulas on your own sheet.

5. Incident and non-conformance write-ups

AI can turn rough notes into a clear, neutral record. Strip every identifier first: no names, lab numbers, ward or bed numbers.

Turn these masked notes into a non-conformance report with headings: what happened, when it was found, immediate action, likely cause, corrective action, preventive action. Use only what is in the notes. Write [NEEDS EVIDENCE] next to anything the notes do not support. Neutral tone, no blame. NOTES: [paste masked notes]

Check: the quality lead confirms every statement against records. The root cause is a decision for your team, not the tool.

6. CPD revision and staff teaching

Explain [topic, for example the main sources of pre-analytical error in clinical chemistry] to a practising laboratory technologist. Plain English. Then ask me eight questions, one at a time, and tell me where my answer is weak.

Check: confirm facts in your textbooks, kit inserts and current guidelines. AI helps you prepare. Your CPD points come from activities recognised under the KMLTTB guideline.

What lab technicians should not do with AI

  1. No result interpretation. Do not ask a public chatbot what a result "means" for a patient, or whether it is critical. Interpretation and critical value calls follow your SOPs and go to the clinician.
  2. No validation or release. An authorised person reviews and releases every report. AI does not sign anything.
  3. No patient details in public tools. Names, lab numbers, dates of birth, results, request forms and photos of forms all stay out. Genetic results need extra care, because genetic data is sensitive personal data.
  4. No slide, gel or plate images into chatbots. Even without a name, an image is patient material. Any image analysis tool is a decision for your lab management and data protection officer.
  5. No changes to reference ranges or methods because AI suggested it. Those follow verified sources and your quality system.
  6. No AI help during registration exams. Follow the Board's exam rules. Use AI to revise before the exam, never during it.

For the wider facility picture, read how hospitals and clinics in Kenya can use AI safely. Colleagues may like our guides for pharmacists and hospital administrators.

How FuKazee trains health teams

FuKazee trains hospital, clinic and health NGO staff to use AI on the paperwork around care, never on clinical decisions. As our AI training for healthcare page explains, every exercise uses made-up patients or fully masked records, and short repeated sessions let one group cover while another trains. Training runs at your facility anywhere in Kenya by arrangement, or live online. We are AI trainers, not laboratory scientists, so your lab manager and data protection officer set the limits. Our certificate of completion is not a government qualification, and we do not claim CPD points for it.

Questions people ask

Can a lab technician in Kenya use ChatGPT at work?

For tasks with no patient information, such as SOP drafts, leaflets from approved text, QC and stock analysis, and revision, we found no KMLTTB rule against it on the pages we checked, as long as your employer allows it. Patient information and anything that interprets a result are the line.

Can AI read a lab result and tell me what it means?

It may give an answer, but you should not use it for a patient. AI can miss context, invent reference ranges or sound sure about something wrong. Interpretation belongs to qualified staff and the clinician, using your SOPs and verified references.

Is it safe to paste QC data into AI?

Internal QC data with no patient link is usually low risk, but check your lab's policy first. Remove anything that links a run to a patient sample. The decision to accept or reject a run stays with you and your supervisor.

Does using AI earn CPD points?

Not by itself. CPD points come from activities set out in the KMLTTB CPD guideline, such as CMEs, workshops and hands-on training. AI can help you prepare for them.

Our lab system has an AI feature. Can we use it?

That is a decision for lab management and your data protection officer, not individual staff. It needs a contract, a check of where data goes, and a qualified person who reviews every output.

Sources

All sources last checked 15 November 2026.

Where FuKazee fits in

FuKazee is a Nairobi training company that has trained more than 3,000 people since 2022. We train health teams to use AI on the writing, sheets and admin around their work, with masked data and clear rules on who decides. Tell us how many lab staff you have and what takes most of their time away from the bench. Email lead@fukazee.com

Want guided support instead?

Not sure which programme fits? Tell us who you are and we will point you to the right path: corporate AI training, the ScholarShift Accelerator for postgraduates, or KCSE study plans and tuition.

Rodney MainaFounder & Director, FuKazee

Rodney has run FuKazee's training since 2022 and still teaches inside every programme line: corporate sessions, the postgraduate cohort and the KCSE examiner webinars. First Class Honours in Journalism and Mass Communication, University of Nairobi.