The IMF says AI touches 40 percent of jobs, and 60 percent in rich economies

The IMF's analysis puts global job exposure to AI at almost 40 percent, and around 60 percent in advanced economies where more work is cognitive. Exposure is not the same as loss. About half of exposed jobs may benefit as AI complements the work; the other half may see tasks or demand displaced.

The IMF's leadership was explicit that the split between benefit and harm is a question of policy and preparation, not a foregone conclusion, and warned AI could widen inequality between and within countries.

Exposure is destiny only if you do nothing. The same job can be helped or hollowed depending on whether the person and the organization build the capability to use AI well. Advanced economies are more exposed, which makes preparation more urgent there, not less.

Exposure is a fork, not a verdict. Preparation decides which branch you take.

The more knowledge-heavy the role or economy, the higher the exposure. White-collar does not mean safe.

Inequality is a live risk. The gains concentrate where capability is built.

Map your roles by exposure, then invest first where exposure is high and capability is low.

Frame AI internally as complement-the-work, and make that real, or displacement fills the vacuum.

Watch the within-company gap, not just the market. Uneven capability widens it fast.

Exposure estimates are broad and contested. Use the direction, not the decimal.

The benefit-or-harm split is not automatic. Assuming benefit without building capability is how you get the harm.