AI and work quality: what controlled experiments show
Is text produced faster with AI also better? Early controlled experiments answer yes — within a precise scope.
An often-cited experiment
Shakked Noy and Whitney Zhang (2023, Science) find professionals doing writing tasks with ChatGPT save time and receive better ratings, with narrower gaps between participants.
What it does not measure
Tasks are short, peer-rated, outside company context. Effects on long, technical or accountable documents are not established.
A complementary benchmark
Fabrizio Dell'Acqua and co-authors (2023, Harvard Business School working paper) show assisted consultants do better on some tasks but worse on one beyond the tool's capabilities. (our executive and employee training programmes)
The limit of common practice
Generalising a gain measured on simple tasks to all of a team's work.
A practical check in Fribourg
Have a colleague blind-review ten documents made with and ten without AI. The perceived quality gap, or its absence, says more than the author's impression. To go further: explore the AI on the Job training in Fribourg, or browse our executive and employee training programmes in Switzerland.
In pictures: AI on the Job in Fribourg
- ai on the job
- Fribourg
- research
- training fribourg