Reading AI summaries is doing something alarming to human memory, study finds
A new study has uncovered the surprising effect that AI summaries have on human memory, with new research suggesting that it can make people less able to remember things accurately.
This comes after a team of researchers from Georgetown University and the University of Washington began investigating what actually happens when people are fed a misleading AI-generated account of an event they had previously witnessed.
The results of the study suggests that the problem with AI summaries could extend beyond simply giving out incorrect information. Instead, an error in an AI summary could have potential impact on a person's own memory of what happened.
Yael Eiger, who is a PhD candidate at the University of Washington, said: “I was struck by how bad the summaries were, even at this stage in AI development.
Advert
“It worries me that police departments may be using video summarization technologies without rigorous testing and without an awareness of how incorrect AI-generated summaries could be.”
Meanwhile, Yoshi Kohno, the McDevitt Chair in Computer Science, Ethics, and Society and professor at Georgetown’s Department of Computer Science and Center for Digital Ethics, added: “This study is part of a growing effort at Georgetown focused on exploring the impact and relationship between AIs and humans, grounded in both psychology and computer science.”
How do incorrect AI summaries impact the human brain?
For the study, the researchers conducted an experiment on 331 participants which involved an animated road accident. The video showed a red car approaching a junction, with participants randomly shown either a stop sign or a yield sign.
The car then turned into the junction and collided with a pedestrian crossing the road. Rather than immediately testing the participants on what they had seen, the researchers waited between 24 and 48 hours, where they were then given a written summary of the video.
Some participants received an accurate account, while others were given a version containing a deliberately altered detail.
The test found that those who had been exposed to the misleading summary were a lot less likely to remember what they had originally seen.
Mattea Sim, an assistant research professor at the Massive Data Institute (MDI) in Georgetown’s McCourt School of Public Policy who led research for the study, explained: “AI is a new method of delivering misinformation, and it has the potential to create these false memories for people who are reading that information.
“I think we should be thinking really deeply and critically about whether and how AI should be used to summarize information, especially in high-stakes settings.”