Can AI governance frameworks reduce the gap between expert optimism and public concern?

Why is there a gap between expert optimism and public concern?

The gap exists because experts and the public have fundamentally different perspectives on AI risks and benefits. A 2018 U.S. public opinion survey found that Americans are deeply fragmented: they recognize AI's promise but express significant concerns about its long-term effects on jobs, legal systems, and national security [5]. Meanwhile, experts often focus on technical innovation and efficiency gains, as seen in the EU AI Act's risk-based approach, which prioritizes market competitiveness and safety standards over public engagement [1]. This disconnect is compounded by the fact that local U.S. policymakers—key players in shaping AI regulation—report feeling underprepared and inadequately informed to make AI-related decisions, according to a 2022-2023 survey of local officials [8]. The same survey found that policymakers anticipate major societal risks like increased surveillance, misinformation, and political polarization, alongside potential benefits [8]. So, the gap is not just about knowledge; it's about divergent priorities and a lack of inclusive dialogue.

How can governance frameworks actually reduce the gap?

Frameworks can bridge the gap by embedding public participation, transparency, and trust-building mechanisms from the start. The EU AI Act establishes a collaborative governance system with a European AI Board, an advisory forum, and a scientific panel, but its design is still top-down and expert-driven [1]. Research on human-centric AI governance argues that true human-centeredness requires moving beyond user-centered design to include community- and society-centered perspectives, enabled by inclusive governance modalities like civic tech and open communication [6]. A comprehensive framework for governing generative AI emphasizes the need for adaptive, participatory, and proactive approaches that align AI development with societal values [2]. Similarly, experts in AI and brain science stress the importance of open discussions that include diverse lay opinions, not just experts, to avoid over-consolidation of power [7]. These studies converge on the same conclusion: frameworks that actively solicit and incorporate public concerns—through advisory forums, public consultations, and transparency requirements—are more likely to reduce the trust gap.

What are the current limitations of AI governance frameworks?

Despite their promise, current frameworks face significant limitations. A bibliometric analysis of research on the EU AI Act found a critical lag between AI technological advancement and the development of policy and regulation, especially for high-risk AI systems [3]. This means frameworks are often reactive, not proactive. Furthermore, the concept of 'human-centric AI' is used ambiguously in policy documents, risking the downplaying of promises for truly emancipatory technology that promotes human wellbeing [6]. In the U.S., local policymakers show partisan divides: Democrats strongly support regulation, while Republicans shifted toward majority support only between 2022 and 2023, highlighting the need for bipartisan coordination [8]. Additionally, China's ambition to lead global AI governance faces geopolitical challenges, limiting its ability to set norms that might bridge expert-public gaps [4]. These limitations suggest that frameworks alone are insufficient; they must be backed by capacity-building initiatives, public education, and sustained political will to be effective.

About These Sources

This answer is built on 8 peer-reviewed studies — published from 2022 to 2025, 4 from 2024 or later, 6 in Q1 journals, collectively cited 355 times — selected as the most relevant from 11 studies that passed quality screening, drawn from 78 papers retrieved from a database of over 500 million.

Sources used in this answer

Related research questions