Media and responsible AI governance: a game-theoretic and LLM analysis

Media acts as a critical substitute for formal AI regulation: The paper finds that investigative journalism and media reporting can function as "soft" regulation when institutional AI governance is absent, making media scrutiny particularly important in regions without established regulatory frameworks.

User trust depends heavily on media recommendations: The research demonstrates that users' willingness to adopt safe AI systems is significantly influenced by what commentators and media outlets recommend, meaning media credibility directly shapes whether people trust and use responsible AI products.

Incentive structures determine whether regulation actually works: The study reveals that effective AI governance doesn't automatically emerge—it requires carefully managing the incentives and costs for high-quality media commentary and regulatory oversight, suggesting that simply creating rules isn't enough without supporting the systems that enforce them.

Strategic behavior among all actors shapes outcomes: Using game theory, the paper shows that AI developers, regulators, users, and media outlets all make strategic decisions based on what others are doing, meaning responsible AI governance requires understanding and influencing these interconnected incentives rather than relying on any single actor.

Different regulatory regimes produce different results: The findings emphasize that the effectiveness of AI governance depends on the broader regulatory environment, suggesting that one-size-fits-all approaches may not work and that policymakers should consider how their chosen regulatory approach influences the behavior of developers, media, and users.