Using AI to enhance societal decision making

Abstract

The rapid advancement toward Artificial General Intelligence (AGI) necessitates a proportional improvement in societal decision-making to navigate increasingly high-stakes scenarios and compressed developmental timelines. Human institutions frequently suffer from epistemic failures and coordination deficits, which specialized AI systems are uniquely positioned to mitigate. By prioritizing the development of “epistemic tools”—such as automated fact-checkers and forecasting systems—and “coordination tools”—including AI-enabled negotiation and verification frameworks—humanity may significantly enhance its ability to manage existential risks. This strategy of differential technology development seeks to accelerate the rollout of safety-promoting capabilities before more dangerous, broadly capable agents emerge. While such initiatives face challenges, including market under-incentivization of non-commercial applications, the risk of inadvertently accelerating general AI R&D, and potential dual-use concerns, targeted interventions by entrepreneurial researchers can mitigate these downsides. Focusing on applications with high pro-social utility and low strategic risk offers a viable path toward institutional resilience. The cultivation of a specialized field dedicated to AI-enhanced decision-making is essential for ensuring that collective wisdom and cooperative capacity keep pace with technological acceleration. – AI-generated abstract.