Summary: Meetings with AI may reduce the likelihood of conspiracy theories being 80 % less. Scientists found that AI provided tailored, fact-based retorts to individuals ‘ crime says, leading to a profound change in their views.
In one out of four cases, members disavowed the plot completely. According to the study, AI has the potential to combat propaganda by getting people to know them directly and personally.
Important Information:
- AI reduced conspiracy beliefs by 20 % after short conversations.
- One in four individuals completely refuted the crime principle they backed down on.
- The result persisted for two weeks after the AI talk.
Origin: American University
Conspiracy theorists often believe that” they’re so far down the rabbit hole of conspiracy theories that they’re lost for good.” This usually accepted belief is then crumbling.
A team of researchers from American University, Massachusetts Institute of Technology, and Cornell University conducted a groundbreaking study to demonstrate how crime scientists ‘ beliefs changed after brief exchanges with artificial intelligence.
Study members believing some of the most deeply entrenched theories, including those about the COVID-19 crisis and fraud in the 2020 U. S. presidential election, showed significant and sustained reductions in crime conviction following the conversations.
Consistent ideas are a major public health problem because they are fueled by fragmentation in politics and fed by propaganda and social press. They frequently cause a rift between scholars and their friends and relatives.  ,
According to the effects of the YouGov survey from December last year, a large portion of Americans believe a variety of secretive lies.
According to Thomas Costello, associate professor of psychology at American University and lead author of the novel study published in the journal , Science, the general consensus that the results problem is that conspiracy theorists agree to their values because they have meaning for their names and because they reflect underlying desires and drives, Costello asserts in the new research.
In fact, the majority of approaches have focused on preventing people from initially believing conspiracies.
” Many conspiracy believers were indeed willing to update their views when given compelling counterevidence,” Costello said.
At first, I was surprised, but reading the conversations changed my mind. The AI was skilled at being amiable and building rapport with the participants, and she was able to provide page-long, highly detailed explanations of why the given conspiracy was false in each round of conversation.
More than 2, 000 self-identified conspiracy believers participated in the study. About 20 % of the average participant’s confidence in their chosen conspiracy theory was affected by the AI conversations, and about 1 in 4 participants, who had all previously believed in the conspiracy, disavowed the conspiracy after the conversation.  ,
Until now, delivering persuasive, factual messages to a large sample of conspiracy theorists in a lab experiment has proved challenging. For one, conspiracy theorists are often highly knowledgeable about the conspiracy—often more so than skeptics. Conspiracy also has a wide range, causing that the evidence supporting a particular theory can vary widely.  ,
AI as an intervention
The new study comes as society debates the benefits and drawbacks of artificial intelligence. Large language models that drive generative AI are potent knowledge reservoirs. The study highlights one way that these reservoirs of knowledge can be used for good: by empowering people with more accurate beliefs, according to researchers.
Artificial intelligence’s ability to connect with various information sources in a matter of seconds allows it to tailor counterarguments to particular conspiracies of a believer in ways that are n’t possible for humans to do.  ,
” This is a major limitation of previous attempts to discredit doubtful beliefs: It’s not a simple task to guess what people’s true beliefs are,” said Gordon Pennycook, associate professor of psychology at Cornell University and co-author of the paper.
” In contrast, the AI can respond directly to people’s specific arguments using strong counterevidence. This provides a unique opportunity to assess how responsive people are to counterindicative evidence.
Researchers created the chatbot to be incredibly persuasive and draw in users to engage in such thought-provoking dialogues. GPT-4, the AI model powering ChatGPT, provided factual rebuttals to participants ‘ conspiratorial claims. Participants were asked to describe a conspiracy theory they believe in and provide supporting evidence in two separate experiments.
Then, the participants had a conversation with an AI. By addressing specific evidence, the AI intended to contest beliefs. Participants in a control group had an unrelated discussion with the AI.
To tailor the conversations, researchers provided the AI with participants ‘ initial statement of belief and the rationale. This setup allowed for a more natural dialogue, with the AI directly addressing a participant’s claims. The conversation averaged 8.4 of the participant’s minutes and involved three rounds of interaction, excluding the initial setup.
Ultimately, both experiments showed a reduction in participants ‘ beliefs in conspiracy theories. Two months later, the researchers assessed the participants and discovered that the effect persisted.
Although the findings are encouraging and point to a future where AI can, when used responsibly, help to lessen conspiracy theories, further research on long-term effects, using various AI models, and utilizing real-world situations will be required.  ,  ,
Our study shows that generative AI can also be a part of the solution, according to David Rand, a paper co-author and professor of the MIT Sloan School of Management. Large language models like GPT4 have the ability to thwart conspiracies on a large scale.
Additionally, members of the public interested in this ongoing work can visit a , website , and try out the intervention for themselves.
About this news about psychology and AI research
Author: Rebecca Basu
Origin: American University
Contact: Rebecca Basu – American University
Image: The image is credited to Neuroscience News
Original Research: The findings will appear in Science
