Generative AI Interactive Coaching Service & Dynamic Simulator

Helping Media Professionals Evaluate Large Language Model (LLM) Outputs with AI Verification and Counter-Narratives Against Disinformation

The Generative AI Interactive Coaching Service & Dynamic Simulator consists of two modules.

The Interactive Coaching and Simulation Module (ICSM) helps media professionals understand the principles of generative AI (e.g., Large Language Models) and their associated risks, focusing on how to construct prompts and interact with systems to produce trustworthy content. It includes guidance on reliability analysis of LLM outputs and verification of their accuracy.

The Counter-narrative Recommendation Module (CNRM) offers on-demand suggestions for responding to disinformation in simulated online conversations, using debunk reports and LLMs for pro-social purposes. It helps users develop counter-disinformation strategies and supports claim verification workflows.

The two modules are interconnected. Users apply their learnings from the ICSM (e.g., prompt construction and reliability analysis) in the CNRM for practical LLM applications, such as analyzing claim veracity and customizing prompts. The ICSM also helps users address inaccuracies in the CNRM’s truthfulness explanations and understand the limitations of LLMs in providing reliable outputs.

The service will integrate the Personal Companion for Understanding Disinformation for identifying logical fallacies and linking to the Trustability and Credibility Assessment Service, providing a holistic approach to media professionals’ disinformation management.