Artificial intelligence was supposed to help employees finish tasks faster, reduce workloads and reclaim valuable time. Instead, new research suggests that people using AI at work could be spending up to 90 minutes longer on the job each day than colleagues who do not use the technology.
An analysis of 9,140 knowledge workers by workplace productivity company ActivTrak found that AI users had average workday spans of 8.4 to 9.1 hours, compared with 7.6 hours among non-users.
The findings raise questions about whether AI is genuinely making workplaces more efficient or whether employees are using the technology to take on additional responsibilities.
AI Users Are Working Longer Days
The ActivTrak Productivity Lab analysed digital work activity involving 9,140 knowledge workers across 467 companies between 1 April and 25 September 2026. Using proprietary workplace activity data rather than employee surveys, researchers examined how AI usage relates to working hours, application switching and uninterrupted focus.
The figures refer to workday spans, not necessarily hours of uninterrupted active work.
ActivTrak divided AI users into different categories based on their daily usage. Heavy users spent more than 45 minutes across at least five sessions using AI tools each day, while regular users spent between 10 and 45 minutes and light users spent less than 10 minutes.
Heavy AI users also recorded an additional 18 minutes of after-hours work compared with non-users.
AI users work up to 90 minutes longer every day than non-users, according to an analysis of more than 9,000 employees published Thursday by productivity software company ActivTrak. pic.twitter.com/Vyfnqz2Nga
— matthew sigel, recovering CFA (@matthew_sigel) November 2, 2026
However, the proportion of their workday spent outside scheduled hours was almost identical, at 32.3% for heavy users and 31.9% for non-users. This suggests that their longer overall working days account for much of the difference in after-hours activity.
They switched between applications an average of 1.1 times per hour, 15.2% more frequently than non-users, who averaged 0.95 switches.
Their uninterrupted focus sessions were also shorter. Heavy users averaged 37.8 minutes of uninterrupted work, compared with 48.2 minutes among light AI users.
'AI may help employees complete tasks faster, but our research shows it contributes to longer workdays, more application switching and shorter focus sessions,' Gabriela Mauch, chief customer officer and head of the ActivTrak Productivity Lab, said.
The way employees access AI tools also appears to influence how frequently they switch between applications. Those using AI embedded in their primary work applications averaged 0.6 switches per hour, compared with between 1.7 and 4.2 switches among employees using browser-based AI tools. ActivTrak describes the additional effort involved in moving between separate applications as a 'swivel-chair tax'.
Mauch offered a possible explanation for the broader productivity paradox, telling Bloomberg: 'We're in an extremely productivity-driven economy, where if we get some, we want more.'
However, the analysis does not establish that AI use causes longer working hours. It is also possible that employees who already work longer days are more likely to adopt AI tools.
AI Could Be Creating More Work
Separate research from the University of California, Berkeley's Haas School of Business offers another perspective on the trend, suggesting that AI could expand employees' responsibilities rather than simply reduce their workloads.
An eight-month ethnographic study at a technology company with 200 employees, based on workplace observations and more than 40 interviews, found that generative AI was expanding workers' responsibilities rather than necessarily freeing up their time.
Researchers Xingqi Maggie Ye and Aruna Ranganathan observed that employees increasingly took on tasks they might previously have avoided or left to colleagues.
'Employees worked at a faster pace, took on a broader scope of tasks, and extended work into more hours of the day, often without being asked to do so,' the researchers wrote.
The researchers also found that AI allowed employees to keep several tasks moving simultaneously. Some ran AI processes in the background while drafting documents, reviewing code or attending meetings, with others using multiple AI agents at once.
However, the Berkeley research examined one technology company rather than a representative cross-section of the workforce, so its findings should not be taken as proof that AI is increasing workloads across all industries.
The Productivity Gains Could Come at a Cost
Ye warned that the apparent productivity boost could become difficult to sustain as employees gradually adjust to higher expectations.
'As task scope expands and multiple AI-assisted workflows run in parallel, the workday becomes denser and more cognitively demanding,' Ye said in an interview with UC Berkeley Haas.
She explained that because the expansion of work can initially feel self-driven and exciting, expectations may gradually reset, turning what was once extra effort into standard performance.
While AI may help employees complete some tasks faster, the Berkeley research suggests that workers can end up taking on additional responsibilities rather than using the technology to reduce their workloads.
The challenge for employers will be ensuring that productivity gains translate into genuinely more efficient working practices rather than simply longer days and heavier workloads.
