You’re Not My Real Boss! Workers Balk at AI Agent Managers
Employees are getting more comfortable leveraging artificial intelligence (AI) tools and agents at work, but they still want clear boundaries and human hierarchies, according to a new report from HR software platform Workday, Inc.
Although 75 percent of the 3,000 business leaders surveyed for the report said they were comfortable working alongside AI agents, only 30 percent said they would feel comfortable taking orders from one. Respondents are also leery of situations where AI makes critical financial decisions (40 percent are comfortable) or operates without direct human oversight (24 percent are comfortable).
This echoes other recent research from Stanford University that showed employees are getting more comfortable working with AI provided it’s helping with mundane, low-stakes, or repetitive tasks and not encroaching on human agency, ZDNet reported.
“Many also express concerns about the impact of agents on critical thinking, human connection, and ethical guardrails—a reminder that this technology, while powerful, is dual-edged,” according to the Workday report, AI Agents are Here—But Don’t Call Them Boss.
The paradox between widespread optimism and real-world concerns goes further.
Overall, 88 percent of survey respondents said AI agents will ease their workload by boosting productivity, but that’s not without concerns—48 percent are concerned that AI agents will increase pressure on employees to work faster to keep up. That could potentially lead to burnout or turnover, other research found.
Meanwhile, the Workday report found that employees expect AI agent use to improve employee growth and development (85 percent), work-life balance (80 percent), and job satisfaction (79 percent). But there are major barriers to adoption, including ethics and governance issues such as bias, data privacy, and legal challenges (44 percent) and security and privacy (39 percent). Nearly a third of survey respondents were very or extremely concerned about potential AI misuse.
Machines are taking over repetitive or data-saturated tasks. Humans are being repositioned into roles that require decision-making skills.
— Security Management (@SecMgmtMag) August 12, 2025
In this evolving landscape, security leaders are asking: Are we replacing the human element or elevating it?https://t.co/CPVdkWQkHY
Employees were more likely to trust AI agents to be fair and unbiased when making decisions about IT infrastructure and technology provisioning, but when it comes to decisions about people, money, and legal compliance, trust shifts to humans.
Respondents were most comfortable with AI agents recommending skills development or areas for improvement, AI helping them complete tasks, and AI agents interacting with each other in the workforce to collaborate and complete work.
“The lasting value of uniquely human skills at work is clear: while AI agents boost efficiency and offer valuable insights, they cannot show empathy, make tough judgment calls, communicate with nuance, or lead people through complex situations the same way humans can,” the report said.
Although workers are not comfortable taking orders from AI agents, respondents noted that the tools still have workforce management benefits, including forecasting and planning (84 percent said AI could benefit their organizations in this area), scheduling and labor optimization (80 percent), and time and attendance (75 percent).
