A recent study published in Nature Human Behaviour reveals that chatbots, specifically large language models like OpenAI's GPT-4, are more persuasive than humans in debates when they leverage personalized information about their opponents. Conducted with 900 participants, the research indicates that while GPT-4's effectiveness matches that of human debaters without background info, having demographic details significantly enhances its persuasive capabilities. This has implications for political campaigns and advertising, raising ethical concerns about the use of AI in influencing public opinion.
"Obviously as soon as people see that you can persuade people more with LLMs, they're going to start using them," says study co-author Francesco Salvi, a computational scientist at the Swiss Federal Technology Institute of Lausanne (EPFL). "I find it both fascinating and terrifying."
Research has already shown that AI can make people change their minds, but it hasn't been clear how persuasive they are compared to humans.
The study measured how participants' views changed after debating with either a human or a chatbot, revealing compelling insights into the persuasion power of LLMs.
The results showed that when debaters lacked background information, GPT-4 was as persuasive as humans, but access to contextual data improved LLM effectiveness.
Collection
[
|
...
]