糖心Vlog破解版

Citation

Balabanova N, Bashir A, Bova P, Buscemi A, Cimpeanu T, Correia Da Fonseca H, Stefano AD, Duong H, Domingos EF, Fernandes AM, Han TA, Krellner M, Ogbo NB, Powers ST, Proverbio D, Santos FP, Ush Z & Song Z (2026) Media and responsible AI governance: A game-theoretic and LLM analysis. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences.

Abstract
This paper investigates the complex interplay between AI developers, regulators, users, and the media in shaping trustworthy AI development. Using evolutionary game theory and large language models (LLMs), we model the strategic interactions among these actors under different regulatory regimes. We explore two key mechanisms for achieving responsible governance, safe AI development, and adoption of safe AI: incentivising effective regulation through media reporting, and conditioning user trust on commentariat recommendations. We show that when high-quality media investigations are sufficiently rewarded and not excessively costly, they can either substitute for weak formal regulation (by directly scrutinising developers) or enhance strong regulation (by monitoring regulators), leading to higher levels of safe AI development and user trust. Our analysis identifies parameter regimes under which full cooperation by all actors-responsible developers, effective regulators, informed media, and discerning users-emerges as a stable equilibrium. Complementary LLM-based simulations broadly corroborate these patterns while also revealing behavioural deviations, particularly among regulator agents, that highlight how real-world decision-makers, and AI models themselves, may depart from idealised game-theoretical predictions. Overall, these results underline that effective AI governance crucially depends on aligning incentives and reducing the costs of rigorous, accurate media scrutiny.

Keywords
AI governance; AI regulation; responsible AI; game theory; LLM; trustworthy AI; behavioural dynamics; media 2