Regarding business decisions, humans must have final say over AI

Artificial Intelligence and other related automated technologies transformed from a distant prospect of the future into an unavoidable aspect of modern life in the span of just two short years. Ever since ChatGPT launched in late 2022 AI has absolutely exploded across virtually all corners of society. 

Automated systems are being incorporated into organizational processes and daily tasks at a rapid rate. While this massive shift in business practices is well intentioned, with an eye toward making employees’ lives easier while increasing productivity, many continue to voice concerns that the lightning fast adoption of AI across industries is too much too soon. 

This issue is perhaps at its most glaring when it comes to decision making. Sure, AI is great for producing poorly written paragraphs within mere minutes, or suggesting related products online shoppers may want to add to their shopping carts, but should humans really be listening blindly to automated recommendations? 

Of course not. AI may one day reach the capacity to match the depth of humanity’s critical thinking skills and pathos, but that day has yet to come – and quite frankly may never. New research just released by the University of Auckland drives this point home by highlighting the importance of always allowing human workers the ability to override AI recommendations and decisions.

Human autonomy over AI is a win-win 

Besides the undeniable fact that current AI models simply don’t match up to highly skilled human professionals, and may end up costing companies large sums of money and countless customers if allowed to operate unchecked, this latest research project tells us that affording human employees the authority to overrule AI decisions immediately helps boost motivation and enhance learning.

After all, if a robot is performing virtually every aspect of a job and making the final decisions – what incentive do human workers have to hone their craft and improve?

“Imagine you’re a financial specialist at a bank,” says Business School lecturer Dr. Frank Ma in a university release. “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.”

Autonomy and incentives pay off in the long run

In collaboration with Stijn Masschelein and Vincent Chong from the University of Western Australia, Dr. Ma put together a series of online assessments encompassing 161 participants. Each of the tasks were intended to simulate real-world scenarios in sectors in which AI is beginning to play a role in decision making.

The ensuing findings indicate that when workers are allowed and empowered to overrule AI decisions early on, they usually develop greater levels of motivation and become better equipped to understand complex tasks. 

Importantly, this project also analyzed the role of incentives. This approach led to the revelation that combining early AI autonomy and pay incentives fosters an especially productive work environment characterized by engaged workers who learn faster.

“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,” Dr. Ma adds. “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.”

Don’t delay human autonomy over AI

In conclusion, the research team posits it’s imperative to provide workers interfacing with AI on a regular basis with tons of flexibility as soon as possible. Why? More autonomy means more chances to overrule automated decisions. Meanwhile, incentive pay helps motivate employees to put in more effort when it comes time to make an accurate final call or big decision.

“Employees with incentive schemes and immediate flexibility get a better understanding of their roles and improve their performance. We believe this is due to developing a more in-depth understanding of how the computer system or AI generates its decision,” Dr. Ma concludes.

Moreover, this study also suggests delaying AI autonomy will likely have a detrimental effect on employee motivation, in all likelihood limiting learning and performance in the long run.