AI could reduce belief in conspiracy theories, study suggests

There may be those who advocate conspiracy theories They were convinced to give up their faith after interacting with an AI-powered chatbotThis is according to a study conducted by psychologists from the Massachusetts Institute of Technology (MIT) and Cornell University in the United States.

According to the article published in Science Around Friday (13), One in four participants in the experiment gave up their initial beliefsOthers were 20% less confident in their theories, on average. 2,190 Americans participated in the survey.

Volunteers He shared his conspiracy ideas with a botThese included topics like the Illuminati and the covid-19 vaccine, as well as different versions of the John Kennedy assassination and the 2020 presidential election. These people had to rate their confidence in such theories on a scale of 0 to 100.

During the interaction, AI provides evidence that challenges conspiracists’ beliefs, confirmed with 99.2% accuracy by a professional fact-checkerThe technology was trained to be empathetic during conversations, avoiding insults and insults to participants.

Artificial intelligence in combating disinformation

The researchers stressed that the result shows how technology can play an important role in combating disinformation, persuading even the most resistant conspiracy theorists to change their minds. Some of these individuals retained their new positions even two months after the study ended..

Besides, They are less likely to follow social media profiles that advocate similar ideas to their own Before joining the conversation with AI, some of the former conspiracists were even more willing to oppose those who advocated such theories.

Its name is “DebunkBot” The engine used in the study was powered by GPT-4It’s the same model that OpenAI’s ChatGPT is based on, and many participants thanked it for helping them see a different side of the topics discussed.