U.S. employees are spending an average of 87 minutes per day — amounting to 47 eight-hour workdays in a year — using and interacting with artificial intelligence tools, according to a new survey.
However, the research, by human resources software platform BambooHR, also revealed that workers spend more time troubleshooting errors and on “prompt iteration” (42% of AI time) than they do on immediately productive activity (35% of AI time). Prompt iteration is the process of improving the instructions given to an AI model to achieve more accurate or higher-quality results.
The other 23% of AI time is spent on exploratory use of AI tools, BambooHR found in the survey of 1,608 full-time, salaried desk workers in the United States. The participants included a subgroup of 520 HR professionals with a title of manager or above. The report was released on Sept. 1, just a couple weeks before Anthropic CEO Dario Amodei publicly called for slower AI development amid fears of serious risks the technology could pose.
The amount of time spent on AI is not spread smoothly among the workforce. The 87 minutes per day average breaks down as 101 minutes for executives (C-suite and vice presidents) and 54 minutes for others.
Even so, many survey participants said they believe junior employees’ reliance on AI limits their growth. In an interesting, counterintuitive twist, the sentiment was more prevalent among the younger workers themselves.
As AI usage grows, the tools are displacing conversations between workers, such that the typical employee is working solo for more than half (56%) of the day. The rest of the worker’s time is spent interacting with others, but 41% of that time still involves AI, while 59% is spent having conversations.
“The data suggests that AI use could be pulling away from time that would otherwise be spent on collegial collaboration,” BambooHR wrote in its survey report. “Workers who express hesitancy around AI spend an extra 15 minutes on human interaction per day compared to their AI-enthusiast counterparts.”
Having some degree of hesitancy to use AI tools is a fairly common trait, affecting 33% of workers. In fact, 86% of workers said their mindset on AI usage has been affected by recent news, such as coverage of data centers, environmental effects and job impacts, the report noted. However, only 3% have actually refused to use AI tools.
Another impact from the spread of AI is a decline in mentorship. Half of workers (49%) said there is a meaningful communication gap between senior and junior employees, and 48% said cross-generational mentorship has become harder as junior workers lean on AI instead of senior colleagues, BambooHR reported.
“The avoidance of getting help from a human looks to be mostly a mix of self-sufficiency and self-consciousness: 53% of workers opt to find an answer themselves before asking anyone, while 27% would rather ask AI than admit to needing help,” the report said.
