Opinion | As AI revolutionises work, get ready for the ‘great compression’
Companies must harness innovations in artificial intelligence to ensure narrowed gaps in productivity don’t undermine quality
When a three-month hire can match a seasoned professional, skill-based hierarchies lose their logic. This “great compression” calls for a fundamental redesign of how institutions operate.
A recent study at a Fortune 500 company illustrates the pattern. AI improved overall performance, but unevenly: the bottom 20 per cent – as defined by the study’s skill index – underwent a productivity increase of more than 35 per cent, the middle group experienced modest gains and the top 20 per cent remained essentially unchanged. More strikingly, learning curves compressed – workers with three months of tenure performed about as well as those with a full year on the job.
In effect, the dispersion of performance shrank: AI pulled the lower tail towards the mean while leaving the top essentially unchanged, creating a narrow band of solid, reliable output.
Another recent study on Japanese taxi drivers shows a similar effect. AI navigation systems that anticipate demand reduced the time spent searching for the next fare. Low-skilled drivers improved by about 7 per cent, while veterans saw little change or even slight declines. The productivity gap between novices and veterans narrowed.