Date context: Senaste öppna författarversionens datum
Brynjolfsson, Li and Raymond study a phased AI-assistant introduction among 5,172 customer-support workers. The current open author version reports 15 percent more resolved issues per hour on average. Less-experienced workers improve speed and quality; more-experienced workers have smaller speed gains and some quality declines. An average hides important differences.
Brynjolfsson, Li and Raymond; The Quarterly Journal of Economics · 2024-11-06
Author-version abstract, version date and licence checked; full independent study review remains outstanding.
What the source shows
Brynjolfsson, Li and Raymond study a phased AI-assistant introduction among 5,172 customer-support workers. The current open author version reports 15 percent more resolved issues per hour on average. Less-experienced workers improve speed and quality; more-experienced workers have smaller speed gains and some quality declines. An average hides important differences. Source: Brynjolfsson, Li and Raymond; The Quarterly Journal of Economics, 2024-11-06
What the material does not establish
This is a particular support system and setting, not every HR task. Do not mix these results with the older working paper's 5,179 workers and roughly 14 percent. The author abstract was checked, not the full analysis independently reproduced. Source: Brynjolfsson, Li and Raymond; The Quarterly Journal of Economics, 2024-11-06
For HR, AI + JJ
AI + JJ, application: Compare the same fictional HR task with and without support, across experience levels. Measure quality and time together. Who learns most, and what should an experienced colleague check? Source: Brynjolfsson, Li and Raymond; The Quarterly Journal of Economics, 2024-11-06