2023-06-30 01:11:59
Published on 30.06.2023
Estimated reading time: less than 1 minute
Technologies » ChatGPT has a special talent for convincingly generating false information, according to a study by the University of Zurich (UZH). Participants had a harder time spotting fake news in tweets produced by the chatbot than in those written by humans.
In addition, the texts produced by the software were deemed to be more understandable. It is therefore a “double-edged sword”, write the authors in the journal Science Advances.
Giovanni Spitale and Federico Germani, postdocs at the Institute of Biomedical Ethics and History of Medicine at UZH, submitted tweets written either by artificial intelligence or by humans to 697 people. Topics ranged from coronavirus to climate change to cancer. atS
1688106806
#ChatGPT #blurs #tracks #Liberté