Every parent asking "which course should my child take?" is really asking a scarier question right now: will this job even exist in 10 years?

Fair question. Here's the honest answer — not the panic, not the hype.

The jobs AI eats first

AI is fastest at anything that's repetitive, rule-based, and screen-based.

• Data entry clerks
• Basic bookkeeping
• Call center scripts (Tier 1 support)
• Junior legal document review
• Basic translation
• Entry-level copywriting (product descriptions, generic blog content)

If a job can be fully described in a checklist, a machine will do the checklist cheaper and faster. That's not opinion — that's just what's already happening in Nairobi offices right now.

The jobs that survive — and why

1. Anything hands-on and physical Electricians, plumbers, welders, mechanics, HVAC techs. A robot can write an essay about wiring a house. It can't currently climb into your ceiling and do it — especially in Kenyan homes where every wiring job is a slightly different improvisation. TVET pathways just got a lot more valuable, not less.

2. Healthcare with a human in the room Nurses, clinical officers, physiotherapists, community health workers. AI can read a scan. It can't hold a scared patient's hand, adapt to a chaotic ward, or make a judgment call in a language and context a model wasn't trained on.

3. Skilled trades that require judgment on-site Construction supervisors, surveyors, agricultural extension officers. Real-world variables — soil, weather, materials, people — don't fit neatly into a prompt.

4. High-trust, high-stakes decision roles Doctors (final call), judges, senior accountants signing off on audits, engineers stamping structural drawings. AI will assist. Someone still has to own the decision and the liability.

5. Jobs built on relationships, not information Sales (real B2B sales, not order-taking), teachers, counselors, HR. The valuable part was never "knowing facts" — it's trust, persuasion, and reading a room.

6. Anyone who can direct AI, not just use it This is the new skill layer, not a job title: people who can brief AI well, judge its output, and fix what it gets wrong. Every field will have a version of this — the accountant who uses AI to do 80% of the work and catches the 20% it botches wins over the one who refuses to touch it.

The real pattern

It's not "creative vs technical" or "degree vs trade." The line is:

Does this job require being physically present, personally accountable, or genuinely trusted by another human?

If yes — you're safe for a long while. If the whole job is producing information on a screen with no judgment call attached — that's the part getting automated first, regardless of how prestigious it currently sounds.

What this means practically

Don't just pick a career because "it's safe from AI." Pick one where AI makes you more dangerous, not obsolete — a nurse who uses AI diagnostics, an electrician who quotes jobs faster with AI, a teacher who personalizes lessons with it. The people who get replaced aren't in the "wrong" field. They're the ones who never learned to work with it.

Comparing courses, colleges, and TVET institutions in Kenya? Explore your options on Elimys — discover, compare, apply.