The AI Workload Creep: We aren’t working less; we are fact-checking more.
AI was supposed to give people time back. The early data says it gave them more to do. Welcome to the invisible workload.
We are suffering from the “Enjoyable experimentation” paradox. It’s easier than ever to rapidly generate drafts, run data queries, and prototype code. But generating the output isn’t the bottleneck—it’s verifying the output.
When the cost of generation approaches zero, the volume of generation approaches infinity. What we’re left with is an entire knowledge economy spending their days functioning as middle-management editors for an over-eager, occasionally-hallucinating junior analyst.
We didn’t escape the workload; we just shifted it from creation to correction.
