Generative AI tools, like ChatGPT, are increasingly being used in the workplace, but getting the interplay right is often crucial to ensuring employees stay on board. For instance, research has shown that AI can often strip away the agency workers so crave at work, resulting in us working to support the AI rather than the other way around.
This is the inverse of the narrative whereby technology liberates us from the mundanities of life that we all loathe and frees us to do the very human things that often make our working lives so enjoyable. Central to this is exploring when AI should take the reins and when we should be in charge.
The power to overrule
A recent study from the University of Auckland suggests that when we give employees the power to overrule an AI companion, it’s crucial that we give that power as early on as possible. The study suggests that when we do this, employees become smarter decision makers and better able to learn from each situation than if the power is given to them later on, when they’ve already had a spell of acting on AI-supported autopilot.
“Imagine you’re a financial specialist at a bank,” the researchers explain. “You input details for a mortgage application, and the AI system recommends declining it. While the system is based on hard data, as a human, you can recognize nuances—’soft’ information—that AI can miss. This is where the ability to overrule the system is beneficial.”
It’s similar to previous studies looking into autonomous vehicles. For instance, Stanford researchers found that we can take some time to regain our bearings if we tune out while our vehicle drives itself. This lack of focus can have deadly consequences if we’re unable to regain control in a safe and efficient manner when the need arises.
Stopping thinking
It’s a finding replicated in a study from ESCP Business School, which explores the likely relationship between human intelligence and machine intelligence as our workplaces become more automated. It finds a sweet spot where AI remains a useful ally but doesn’t do so much of the thinking that we become unable to think for ourselves.
The Auckland researchers conducted a range of tasks that each tried to simulate a number of real-world scenarios whereby AI would be tasked with supporting human decision-making. The analysis found that giving employees the power to override AI early on ensures they’re not only more motivated but also better equipped to complete a range of complex tasks.
Giving autonomy
The researchers also looked at whether incentive pay can help to ensure employees work with AI in the right way. The results suggest that combining early autonomy with incentive pay can indeed help employees learn and become smarter in their work.
“Whether and when to override AI decisions is already a big issue in industries including banking and manufacturing, and it’s going to become one in many others that use algorithms for managerial decision-making,” the researchers explain.
“Overall, our study shows that giving employees the power to have the final say over AI early on is critical to their learning. Humans can pick up on nuances that artificial intelligence can’t, so people need the power to make the final call.”
Flexibility is key
The study suggests that giving employees flexibility and autonomy as soon as possible is key, even if that means they’re overriding their AI companions. Incentive pay can help to embolden employees with the confidence needed to do this, which in turn encourages them to make sure that the decisions they make are as wise as possible.
“Employees with incentive schemes and immediate flexibility get a better understanding of their roles and improve their performance,” the authors conclude. “We believe this is due to developing a more in-depth understanding of how the computer system or AI generates its decision.”
Of course, autonomy also has a wider range of benefits at work beyond making the man/machine interface smarter. For instance, research from the University of Melbourne finds that that giving employees autonomy over their work results in higher performance levels and greater loyalty towards the employer.
At a time when both retaining talent has never been more important and ensuring that people are effectively able to work alongside new technologies, it would seem as though giving autonomy to employees is a great way to achieve both.
