Descriptions:
A new study from Berkeley and Harvard Business School researchers Aruna Rananthan and Shingchi Maggi challenges one of the central narratives around AI productivity: that it will reduce how much people work. Published in Harvard Business Review and based on an eight-month embedded study of a 200-person technology company (April through December), the research found that AI is intensifying work, not reducing it — and that this happens through three distinct and well-documented mechanisms.
The first is task expansion: because AI lowers the barrier to unfamiliar work, employees routinely stepped into responsibilities they would previously have outsourced or avoided. Product managers began writing code; researchers took on engineering tasks. The second is boundary erosion: because starting a task with AI is so frictionless, workers began prompting during lunch breaks, between meetings, and in the moments before leaving their desks — eroding the recovery value of downtime without any explicit employer pressure. The third is parallel multitasking: workers ran multiple agents simultaneously and revived long-deferred projects because AI could handle background processing.
The episode frames these findings as neither strictly good nor bad. Organizations that use AI to expand output rather than simply cut headcount will likely outperform those that treat it purely as a cost-reduction tool. But the researchers also flag real risks: engineers increasingly spent time reviewing and correcting AI-assisted work from colleagues, and many workers only recognized in hindsight that their sense of recovery had quietly disappeared as ambient prompting became habitual. The host connects the findings to the broader agentic shift underway in early 2026.
📺 Source: The AI Daily Brief: Artificial Intelligence News · Published February 11, 2026
🏷️ Format: Deep Dive
