The AI Talent Race Still Needs Actual People
A Renaissance master guides two expressive apprentices through a laptop lesson in a Florentine workshop. Made by Edward Frank Morris
The AI race has acquired so many metaphors about engines that one could forget the people expected to drive them. Use ChatGPT to map your team's actual capability gaps before commissioning another training session in which everyone learns to ask for a cheerful email.
Business Insider reported on 16 October 2026 that Microsoft president Brad Smith warned US immigration restrictions could drive AI talent abroad. That is a policy argument. For an employer reading the story, it also raises a closer question. Where does your own AI capability reside?
If the answer is one enthusiastic colleague who occasionally rescues everyone else's spreadsheets, you have a dependency wearing a lanyard.
Audit the work before the workforce
Choose a recurring business process and describe the stages where judgement matters. Assess tasks rather than ranking employees from a handful of messages. Someone who writes beautifully about AI may still be quite unable to recognise when it has invented a customer policy.
Give the team a realistic task with an agreed definition of success.
Ask participants to explain how they checked the output.
Record where source quality or permissions caused problems.
Repeat the task after focused practice and compare the evidence.
Turn the gaps into practice
Using this anonymised workflow and task-level observations, create a four-week AI practice plan. Map each skill gap to an exercise and observable pass criteria. Include checking sources and recognising when human review is required. Do not infer individual ability or personal characteristics from job titles. Separate training needs from missing tools or unclear processes.
Have the manager and the people doing the work review the plan. A training problem and a broken approval process can look remarkably alike from a distance. Only one is repaired by another webinar.
Our earlier article on AI safeguards explores the wider readiness question.
Measure whether people can complete and verify useful work. Attendance certificates are pleasant. So are hotel mints. Neither proves competence.
Comments