By Christian J Meier
The chatbot ChatGPT can “lie” credibly. The AI often claims false facts that sound plausible. But that’s not all: According to a recent study, the AI can be misused to spread disinformation on social networks: the language model GPT-3 from the Californian company OpenAI, a previous version of ChatGPT, generates false information on command that is more credible for the audience than written by humans. Researchers led by Giovanni Spital from the University of Zurich are now showing this in the magazine Science Advances.