The Intelligence from The Economist Bot the difference: AI’s absence in economic data
46 snips
Feb 27, 2026 
Alex Domash, an economics correspondent who studies productivity and AI adoption, outlines why macro data so far shows little AI impact. Ọrẹ Ogunbiyi, an Africa reporter on regional security, explains a brutal Nigerian attack and the murky interplay of jihadists and bandits. Jon Fasman, a culture correspondent, tells the offbeat life of Maine lobsterman Virginia Oliver and her fierce independence.
AI Snips
Chapters
Transcript
Episode notes
Employment Shortfalls Inflated Productivity Signals
- 2025 showed a puzzling gap: strong real GDP growth alongside sluggish employment growth.
- Domash explains tight US immigration and falling temporary-worker participation removed lower‑productivity workers, exaggerating output-per-worker measures.
AI Adoption Is Widespread But Daily Use Remains Limited
- Roughly 40% of working‑age Americans use AI at work, but intensity is low.
- Studies Domash cites find only ~13% use AI daily and average use is about two hours per week (≈5–6% of working hours).
Measured Worker Efficiency Gains From AI Range 15 to 30 Percent
- Controlled studies show AI users can be 15–30% more efficient on tasks.
- Domash synthesises experiments across consulting, legal, and writing tasks to arrive at that efficiency range.