Scientists Explore the Possibilities of Human-AI Coworking

A human and AI together could be more accurate than either one acting separately.

AI is great for eliminating human error, but humans still have the advantage for many types of decision making. Scientists at Oak Ridge National Laboratory (ORNL), University of Tennessee, Knoxville, and Cheng Kung University in Taiwan, have developed a system for human/AI collaboration that delivers improved performance for some types of decisions.

The method, which is described in Nature Portfolio Journal Computational Materials, calls for the AI to display preliminary observations on the data for human review. According to the press release, “Researchers vote on data, telling the AI to show similar information or change direction, akin to a streaming service generating suggested content based on users’ activity. After initial guidance, algorithms improve to illuminate relevant data with little human input.”

This “active recommender” system improves accuracy, but the system is understandably slower than an AI working alone. For additional details on the technique, see the article in npj Computational Materials.