A recent study carried out by researchers from the University of Auckland and the University of Western Australia has shown that giving employees early autonomy to overrule AI decisions significantly improves motivation and learning in the workplace. The findings suggest that businesses adopting artificial intelligence for decision-making processes should consider empowering employees from the outset to have the final say, rather than delaying this control.
Led by Dr. Frank Ma from the University of Auckland Business School, alongside Stijn Masschelein and Vincent Chong from the University of Western Australia, the study aimed to explore how autonomy over AI impacts employee performance. The research, involving 161 participants, simulated real-world scenarios in sectors where AI is used for managerial decision-making. Participants were tasked with making decisions similar to those in banking and finance, where AI systems often recommend actions based on data.
The study’s key finding was that employees who were empowered to overrule AI from the start felt more motivated and became better equipped to handle complex tasks. Dr. Ma highlighted how human intuition can pick up on “soft” information that AI might miss. For instance, in a scenario where a bank’s AI system recommends declining a mortgage application based on the data provided, a financial specialist might recognise other relevant details that the AI cannot incorporate into its decision.
“Humans can pick up on nuances that artificial intelligence can’t,” said Dr. Ma. “Overall, our study shows that giving employees the power to have the final say over AI early on is critical to their learning.”
Conversely, delaying having autonomy over AI had a negative effect on employee motivation. The researchers observed that workers who were not given early flexibility tended to feel less motivated, limiting their ability to learn effectively and reducing their overall performance.
The study also found that combining early autonomy with incentive pay further improved employee engagement. Employees given the freedom to override AI decisions, coupled with financial incentives, learned faster and were more motivated to perform accurately. According to the researchers, this combination encouraged workers to invest more effort in understanding how AI generates its decisions and in making the final call correctly.
“Employees with immediate flexibility and incentive schemes get a better understanding of their roles,” Dr. Ma noted. “We believe this deeper understanding stems from learning how AI systems process data, which allows them to make more informed decisions.”
The findings carry important implications for industries increasingly relying on AI for decision-making. Sectors such as banking, manufacturing, and finance are already grappling with questions concerning when and how much control employees should have over AI. Dr. Ma emphasised that while AI can process large amounts of data efficiently, it is ultimately human expertise and judgment that can capture details missed by algorithms.
The U.S. Department of Labor has also recommended that companies implementing AI tools do so with transparency and clear oversight. Ensuring that employees are involved in the AI process from the start can safeguard workers’ interests and improve overall outcomes.
As AI technology continues to permeate our professional lives, the researchers argue that businesses should integrate human decision-making with AI systems early on. The study, titled Incentive Contracts and the Timing to Introduce Flexibility on Employee Learning, won the Best Paper Award at the 2024 Accounting and Finance Association of Australia and New Zealand conference.
