What Do Human and Machine Minds Have to Offer Each Other?

In a special issue of Current Directions in Psychological Science (Vol. 35, Issue 3, June 2026), former Editor-in-Chief and APS Fellow Robert Goldstone (Indiana University) pulls together a collection featuring voices at the intersection of psychology and AI.

Public discussion about AI often frames the technology as a rival to human intelligence—something that will eventually outperform and perhaps replace humans. The researchers in this issue largely reject that framing. Cleotilde Gonzalez (Carnegie Mellon University) and Tailia Malloy (University of Luxembourg) argue for complementary intelligence: designing AI systems that work with human cognition rather than against it or around it. Katherine Collins (Massachusetts Institute of Technology) and her colleagues push this discussion further, imagining what it would look like for humans and AI to form genuine intellectual partnerships that grow more useful and more nuanced over time.

Another recurring theme in this collection is that researchers often ask the wrong questions when evaluating AI. Melanie Mitchell (Santa Fe Institute) makes this case through two case studies examining how AI systems handle abstraction and analogy. She argues that AI systems should be evaluated beyond merely passing a benchmark test. Researchers should also consider how AI systems handle variations of a task it has never seen before, and whether they can understand why it succeeds or fails.

Gordon Pennycook (Cornell University) and his coauthors add to the conversation by asking how AI can help us better understand human intelligence. Technology, they note, has always reshaped how we understand our own minds. “A consistent pattern emerges from the history of psychology: Technological advances change the way that we understand ourselves,” they wrote. Judith Fan (Stanford University) highlights generative AI as a potent tool for exploring how humans create new ideas, like pictures, stories, and musical compositions. Fan uses drawing as a case study to show how much cognitive science has left unexplored by avoiding open-ended tasks.

What could go wrong when psychology and AI are combined? Iris van Rooij and Olivia Guest from Radboud University round out the issue by arguing that neither field is on especially solid ground and that combining two shaky foundations doesn’t make a stable one but, rather, creates “a perfect storm.”

The issue is now available online.

Feedback on this article? Email [email protected] or login to comment.