For global professionals, the discovery signals a fundamental shift in policy communication, where the credibility of AI systems—not human experts—may become the dominant force in shaping citizen behavior, with profound implications for governance and democratic processes worldwide.
Chinese scientists have uncovered a striking dynamic in the era of artificial intelligence: large language models (LLMs) are proving more effective than human experts at winning public support for controversial government policies. In a pair of pre-registered survey experiments published in the Policy Studies Journal, researchers tested how endorsements from LLMs—particularly Chinese-developed models—influenced citizens’ willingness to comply with two contentious policies in China.
The results were definitive. While expert endorsements failed to move the needle, AI-generated endorsements significantly boosted public support. This is not merely a matter of technological novelty; the study’s exploratory analysis suggests that the mechanism at work is a perceived improvement in the scientific rigor of policymaking itself. Citizens appear to trust the output of these models as more objective, data-driven, and rational than traditional human authority.
The broader significance for China—and for the world—is profound. As AI becomes embedded in public administration, the locus of authority is shifting from credentialed experts to algorithmic systems. This challenges long-held assumptions about how societies build consensus and enforce compliance. For China, which is aggressively advancing its AI capabilities, the findings suggest a powerful new tool for policy communication, one that could bypass the erosion of trust in traditional institutions. For global observers, it raises urgent questions about the nature of consent, the role of AI in governance, and the ethical boundaries of machine-led persuasion.
Why it matters:
This finding upends conventional wisdom about policy communication, revealing that AI models—not human experts—may hold the key to public compliance. For researchers and strategists in communications technology, it signals a new frontier where algorithmic credibility becomes a critical variable in governance and social engineering.
ScientificChina — tracking what’s happening in Chinese science, technology, research, and industrial innovation in a way global professionals can actually use.
Follow ScientificChina for deeper insight into China’s evolving science, technology, and industrial landscape.
No products in the cart.