Shadow AI is already inside your organisation. The only open question is whether leadership governs it or discovers it later, at cost. That is the argument of my Business Daily column on concealed AI use — and events since publication have only sharpened it.
The precedent is familiar. A decade ago employees brought Dropbox, Google Docs and Slack into the workplace before anyone approved them — not out of defiance, but because official tools were too slow, clunky or absent. Shadow AI — the unsanctioned use of AI tools by employees without employer oversight — is the same phenomenon at far greater velocity, in Kenyan workplaces as much as anywhere on the continent.
The risks fall into three categories. One, data exposure: client records and commercial information fed into unprotected public tools. Two, defective output: biased or inaccurate results carrying legal exposure when they touch hiring or lending decisions. Three, cultural erosion: when staff conceal their productivity tools from their managers, trust decays in both directions.
The governance response is practical, not punitive: validate the need the tools are meeting, set clear rules on authorised platforms and data handling, train for bias and privacy awareness, procure enterprise-grade alternatives, and assess usage periodically. With the EU's AI Act partially in force and Kenya's Artificial Intelligence Bill following the same trajectory, African organisations that monitor shadow AI today will meet enforcement tomorrow already compliant — and will have converted a hidden risk into a working advantage.
Read the full column at Business Daily Africa