Your AI Upskilling Budget Is Probably Aimed at the Wrong People

A sweeping IMF study of 165 million U.S. jobs just upended the standard playbook for AI workforce strategy — and the findings should make every CHRO rethink how they're spending their L&D budget. The research introduces what the IMF calls the "complementarity divide": AI doesn't uniformly create or destroy jobs. It rewards roles where humans and AI genuinely augment each other — and penalizes roles where AI can simply replace the human contribution. In fact, occupations with heavy AI exposure but low human-AI complementarity saw a 3.6% drop in employment, even in the tightest labor markets. There's also a stark skill-depth gap that most corporate training programs are ignoring. Workers with AI-developer skills — the ability to build, fine-tune, and direct AI systems — command a 7.5–8% wage premium. Workers who just use AI tools earn around 2%. That's a 4x difference, and most company AI training programs are firmly in the 2% camp. In this episode, we break down the IMF framework, explain who's really at risk (hint: entry-level and middle-skilled workers in substitution-prone roles), and walk through a practical complementarity audit that HR leaders can run this quarter to get ahead of the curve.