AI-Generated Content Can Plant False Memories in Humans
Research reveals that artificial intelligence-generated photos and texts can mislead human memory by planting false memories, with individuals showing overconfidence in these erroneous recollections.
Researchers state that artificial intelligence-generated visuals and summaries successfully disrupt people's own memories, leading individuals to misremember original images and display excessive confidence in their misleading memories.
MIT scientist Pattie Maes stated that within the scope of the conducted research, participants were shown a series of photographs, and following a break, some participants were presented with the original photos while others were shown versions modified by artificial intelligence. It was observed that individuals who saw the AI-modified images were more likely to misremember the original visual and exhibited excessive confidence in their false memories.
Another study conducted last year showed that both deepfakes and AI texts planted false memories about entirely fabricated political scandals and reduced voting intentions. On the other hand, while institutions such as governments, banks, and hospitals deploy artificial intelligence, they enlist human experts to monitor erroneous outputs; however, research reveals that users struggle to detect these AI hallucinations and even experts' judgments can be influenced.