Enterprise
AI in HR Decisions Risks 'Efficiency Trap,' Researcher Warns
Automated systems can execute policies consistently but lack the flexibility to handle nuanced employee situations during crises.
Key takeaways
- Australian companies including NAB and Atlassian have announced thousands of job cuts amid increasing workplace automation.
- Agentic AI systems can independently execute HR tasks and implement decisions with limited human supervision, raising concerns about inflexible decision-making.
- Research identifies an 'efficiency trap' where automated HR processes become too rigid to handle unexpected circumstances or employee contexts.
- Human managers provide critical flexibility to question decisions and make exceptions when standard processes no longer fit the situation.
- Researchers propose matching human oversight levels to decision risks rather than eliminating AI from HR entirely.
As Australian companies announce major workforce reductions—National Australia Bank cutting 2,500 jobs, Atlassian eliminating 1,600 positions globally—questions are mounting about artificial intelligence's expanding role in employment decisions.
While organizations aren't publicly stating that AI determines who gets laid off, the convergence of mass restructuring and advancing automation technology has prompted fresh examination of where machines should stop and human judgment must begin.
The efficiency trap
New research from Macquarie Business School identifies what Dr. Miles Yang calls the "efficiency trap" in human resources automation. Organizations pursuing consistent, scalable HR processes through AI risk losing the flexibility required when circumstances shift unexpectedly.
The concern centers on agentic AI—systems that don't merely analyze data but independently execute tasks and implement decisions with minimal human oversight. In HR contexts, these systems could identify recruitment candidates, allocate shifts, monitor performance, or trigger employment actions based on predefined criteria.
"AI can make HR processes faster, cheaper and more consistent, but managing people isn't simply about following a set of rules," Yang said in findings published in Human Resource Management. "When decisions affect someone's career, livelihood or well-being, context and human judgment matter enormously."
Consider an AI system flagging underperforming employees by analyzing productivity metrics, attendance records, and performance indicators. The system might identify someone whose output has declined—but miss that the employee recently assumed caregiving responsibilities, mentored struggling colleagues, or contributed valuable work the algorithm wasn't designed to measure.
A human manager could recognize these circumstances and adjust the response accordingly. An automated process would likely proceed toward its programmed outcome regardless.
Why it matters
The research arrives as companies face pressure to cut costs while managing larger, more distributed workforces. The temptation to automate HR decisions is strong—AI doesn't tire, processes vast information volumes, and applies rules uniformly across thousands of employees. But Yang's work suggests that consistency isn't equivalent to fairness or sound management, particularly during economic shocks, workplace disputes, or unexpected crises when rigid processes can produce decisions that damage both employees and organizational performance.
Finding the balance
Yang doesn't advocate removing AI from HR departments entirely. Routine administrative tasks and straightforward scheduling decisions can benefit from automation. But decisions involving hiring, termination, workplace conflicts, and employee welfare demand greater scrutiny—especially when circumstances are complex or consequences significant.
The research proposes a governance framework matching human oversight levels to the risks and uncertainty involved in each decision type. Critically, managers need both authority and confidence to challenge or override automated recommendations.
"AI should support better decisions about people, not remove the responsibility to make those decisions thoughtfully," Yang said. "The more significant the consequences for an employee, the more important it is that someone can consider the circumstances beyond what the system has been programmed to measure."
The findings were first reported by Phys.org, based on Yang's paper in Human Resource Management.