Can quantum computing think like a human?

It’s no secret that technology has its flaws. One that comes up time and time again is that technology thinks like, well…technology, not like a human.

But in new findings, scientists at College Park, Maryland’s IonQ have demonstrated cognitive decision-making with quantum hardware. In other words, they’re using quantum to help technology make decisions more like a human would.

Understanding this current development, IonQ Senior Principal Scientist Dominic Widdows said, all goes back to psychology from the 1960s and 1970s, which found that the way we develop probability and even technology isn’t quite compatible with the human brain.

“It is rooted in the basic problem that classical logic and classical probability do not describe all of human decision-making very well at all,” Widdows told Technical.ly.

Human decision-making works like this: If you ask a human a question that they don’t know the answer to, Widdows said, often they will form an opinion on the subject as they’re forming an answer based on logic or prior knowledge. From there, humans are prone to stick to that opinion (if you’ve ever gotten into a fight on the internet with someone, that’s this process at work). Another key trait: humans don’t form opinions independently, and often base future opinions on current ones.

That can even be true in the span of a single survey, Widdows noted, citing an experiment that found that respondents would change their answer on whether or not Bill Clinton was trustworthy depending on if they got the question before or after asking if they found Al Gore trustworthy. That related answer system, he said, can be modeled well naturally in a quantum computing system.

Humans, he noted, are also different levels of risk-averse. But many will change a decision based on whether or not they know part of the outcome.

“We’re not actively constructing each branch of what might happen, we’ve got just some vague notion of the future and somehow the different notions of the future are interfering with one another and producing a different kind of decision-making,” Widdows said. “And that different kind of decision-making has been modeled very well using quantum probability.”

So, what are the practical applications of using quantum in this way? It could mean better decision-making models for election results, help with disrupting unconscious bias or better-personalized ads.

But Widdows thinks one possibility is in generative AI, which has spiked in popularity this year. So far, he said there’s still a lot we don’t know “under the hood” about generative AI; quantum, so far, has done well with its model of small amounts of data for image processing and classification.

Combining this with human-based decision-making that brings in additional context and builds on previous answers could yield better results than current generative AI tech and help us understand the tech models more.

“As AI models become even more important in the way we do business but we need to understand them, even better … There’s definitely a story for some of these quantum insights there and the light they already shine on some of our human behaviors,” Widdows said.