AI is more effective at convincing people than humans are.

Certainly! Here’s the revised content with neutral branding:

—

Their findings contribute to a growing body of research showcasing the persuasive capabilities of LLMs. The authors caution that these AI tools can construct sophisticated, convincing arguments with even minimal information about the individuals they engage with. This research has been published in the journal Nature Human Behavior.

“Policymakers and online platforms should seriously consider the threat of coordinated AI-based disinformation campaigns, as we have clearly reached the technological level where it is possible to create a network of LLM-based automated accounts able to strategically nudge public opinion in one direction,” says Riccardo Gallotti, an interdisciplinary physicist at Fondazione Bruno Kessler in Italy, who contributed to the project.

“These bots could be utilized to spread disinformation, and this kind of diffuse influence would be very difficult to debunk in real time,” he adds. The researchers recruited 900 individuals based in the US and collected personal information such as gender, age, ethnicity, education level, employment status, and political affiliation.

Participants were then paired with either another human opponent or GPT-4 and tasked with debating one of 30 randomly assigned topics—such as whether the US should ban fossil fuels or whether students should wear school uniforms—for a duration of 10 minutes. Each participant was instructed to argue either in favor or against the topic, and in some cases, they received personal information about their opponent to better tailor their argument. At the end, participants expressed how much they agreed with the proposition and whether they believed they were debating with a human or an AI.

—

Feel free to let me know if you need further adjustments!