Modeling the dynamics of misinformation spread: a multi-scenario analysis incorporating user awareness and generative AI impact

The proliferation of misinformation on social media threatens public trust, public health, and democratic processes. We propose three models that analyze fake news propagation and evaluate intervention strategies. Grounded in epidemiological dynamics, the models include: (1) a baseline Awareness Spread Model (ASM), (2) an Extended Model with fact-checking (EM), and (3) a Generative AI-Influenced Spread model (GIFS). Each incorporates user behavior, platform-specific dynamics, and cognitive biases such as confirmation bias and emotional contagion. We simulate six distinct scenarios: (1) Accurate Content Environment, (2) Peer Network Dynamics, (3) Emotional Engagement, (4) Belief Alignment, (5) Source Trust, and (6) Platform Intervention. All models converge to a single, stable equilibrium. Sensitivity analysis across key parameters confirms model robustness and generalizability. In the ASM, forwarding rates were lowest in scenarios 1, 4, and 6 (1.47%, 3.41%, 2.95%) and significantly higher in 2, 3, and 5 (19.67%, 56.52%, 29.47%). The EM showed that fact-checking reduced spread to as low as 0.73%, with scenario-based variation from 1.16 to 17.47%. The GIFS model revealed that generative AI amplified spread by 5.7%–37.8%, depending on context. ASM highlights the importance of awareness; EM demonstrates the effectiveness of fact-checking mechanisms; GIFS underscores the amplifying impact of generative AI tools. Early intervention, coupled with targeted platform moderation (scenarios 1, 4, 6), consistently yields the lowest misinformation spread, while emotionally resonant content (scenario 3) consistently drives the highest propagation.