Project Description
The project will be split between 3 students who will be responsible for a part of the project (monitoring one Youtube channel each). Main objectives of the research are the following: 1. Develop and validate a scalable LLM agents; 2. Test effectiveness of counter-narratives generated by LLM agents using F–M–F–F (Fact-Myth-Fallacy-Fact)structure. The study’s plan is simple: We use AI to create a standardized counter-argument to three different disinformation Youtube channels. We post those counter-narratives in the channels comments, and for a week, we track the survival rate of our counter-arguments. This shows us how much the platform and the channel's audience resist corrective information.
Tasks and Responsibilites
Student will monitor one youtube channel from a specific influencer, and 1. select a false claim that already was debunked. 2. Set up and test n8n program and train the AI to do two jobs: first, to recognize the four narrative parts, andsecond, to write the perfect four-part Fact-Myth-Fallacy-Fact (F–M–F–F) counter-comment. 3. For several days, watch what happens to the comment. 4. Keep a strict record of everything that happens to the comment, especially if and when it gets removed or receives backlash. 5. Analyze and report the results