Your Staff Are AI-Familiar, Not AI-Literate

Near-universal use has been mistaken for capability, but using a tool and understanding it are two very different things.

Sign up for my Substack daily AI newsletter here.

See my AI Training course portfolio for corporate Business Leaders here.

Article link: https://open.substack.com/pub/johanosteyn/p/your-staff-are-ai-familiar-not-ai?r=73gqa&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true

Follow me on LinkedIn: https://www.linkedin.com/in/johanosteyn/

AI has already entered the workplace. In many organisations, the problem is that it arrived before the training did. Employees are using it, managers are seeing the output, and boards are approving the spend, but the literacy needed to judge what the tool is actually doing has not kept pace. High adoption has been quietly mistaken for high capability, and they are not the same thing.

CONTEXT AND BACKGROUND

Adoption is already high in many settings. In the education sector, where the pattern shows up early and clearly, Microsoft’s 2026 AI in Education report found that 92 percent of students and education leaders have used AI for school, yet 77 percent of students and 53 percent of educators say they have had no formal AI training at all. The workplace mirrors this. People are reaching for powerful tools daily and learning them by intuition and trial-and-error, not instruction. The result is a workforce that is AI-familiar, fluent in the motions of prompting, without being AI-literate, able to judge, verify and safely apply what comes back.

INSIGHT AND ANALYSIS

Literacy is the part that does not arrive in the box. It means knowing what a tool can and cannot do, recognising a confident hallucination, understanding the data and confidentiality risk in a given prompt, knowing when not to use AI at all, and being able to verify and improve the output rather than forward it untouched. Without those judgements, employees become overconfident users rather than effective ones, and the organisation gets more volume without more reliability. It is the gap between familiarity and literacy where errors, rework and quiet risk accumulate.

In South Africa, this collides with a labour market already under strain, in what Mercer’s 2026 Global Talent Trends research calls a talent paradox. Local firms are accelerating AI for productivity, with 68 percent of executives expecting a high return, while talent scarcity and a trust deficit pull the other way, and only 41 percent of employees report thriving, below the global average. The striking part is the appetite. More than half of local employees fear their skills will soon be obsolete, and 65 percent say they would trade a 10 percent pay rise for the chance to build AI and digital skills. The demand for literacy is coming from below; the supply is stalling above.

Nor can leaders outsource the fix to the education system. According to Salesforce, reported by TechCabal, South Africa’s qualification frameworks typically run on roughly five-year cycles while AI skills change in months, so graduates can be behind before their first day, against a youth unemployment rate Stats SA puts above 60 percent.

I have previously written about this institutional lag, in a piece on how the people teaching the future of business are too often stuck in its past, incentivised to publish research that takes years while the field reinvents itself quarterly. The literacy layer, it turns out, is now the employer’s to build.

IMPLICATIONS

This lifts the skills gap out of human resources and onto the board agenda, because low literacy is a governance exposure, not a training inconvenience. A workforce that cannot tell a sound output from a plausible-looking wrong one produces poor decisions, leaks confidential data into tools no one assessed, places false confidence in AI-generated work, and quietly widens the gap between trained and untrained teams. It also undermines the very return the board approved the spend to capture, since untrained users generate the rework that erases the time AI was meant to save. A firm that has measured adoption but never defined, baselined or owned literacy is not managing the risk. It is assuming it away.

CLOSING TAKEAWAY

The honest position is that adoption is good and literacy is what turns it into value while keeping the risk in check. That means treating AI literacy as a defined capability rather than a vague aspiration: role-based training tied to real workflows, clear standards for verifying output, sensible rules on when AI should not be used, and continuous learning in the flow of work rather than a single onboarding video. The workforce has already shown it wants this. The question is whether leadership will build the literacy layer deliberately, or keep mistaking a busy workforce for a capable one.

Johan Steyn is a prominent AI thought leader, speaker, and author with a deep understanding of artificial intelligence’s impact on business and society. He is passionate about ethical AI development and its role in shaping a better future. Find out more about Johan’s work at https://www.aiforbusiness.net