Humans and AI: A good team? Study calls for clear task division

According to de Volkskrant, a new study by the Massachusetts Institute of Technology (MIT) shows that collaboration between humans and AI often falls short of expectations. Instead of achieving better results, human-AI teams tend to perform worse compared to qualified individuals or AI alone. For instance, when AI and humans were tasked with classifying birds, AI achieved a 73 percent accuracy rate, while humans scored 81 percent. However, the collaboration between the two resulted in a higher score of 90 percent, marking a rare instance of "human-AI synergy."

AI expert Eric Postma (Tilburg University) argues that this study highlights the importance of clear task division between humans and AI. When roles are well-defined, humans and machines can excel together, particularly in creative tasks. However, for more straightforward tasks, they seem to perform better individually.

Read the full article in the Volkskrant here.