Expert Survey: AI Reliability&Security Research Priorities

2025-05-27

AI Summary

The AI reliability and safety community lacks empirically grounded evidence to guide research prioritization. Method: To address this gap, we conducted a systematic Delphi study involving 53 domain experts, covering 105 research directions aligned with general AI safety requirements. We developed the first comprehensive taxonomy of AI safety research, integrating multidimensional classification, weighted consensus aggregation, and impact-potential modeling to enable data-driven, quantitative prioritization. Contribution/Results: This work fills a critical empirical void in AI safety research prioritization by delivering an openly reusable, expert-consensus-based ranking of research direction influence. It provides the first large-scale, evidence-informed decision-support framework for global AI R&D strategic investment and resource allocation, grounded in rigorous expert elicitation and quantitative modeling.

Technology Category

• Philosophy and Ethics of AI: Safety, Robustness & Trustworthiness
• Humans and AI: Learning Human Values and Preferences
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Security and Privacy: Large-scale security measurements

Abstract

Our survey of 53 specialists across 105 AI reliability and security research areas identifies the most promising research prospects to guide strategic AI R&D investment. As companies are seeking to develop AI systems with broadly human-level capabilities, research on reliability and security is urgently needed to ensure AI's benefits can be safely and broadly realized and prevent severe harms. This study is the first to quantify expert priorities across a comprehensive taxonomy of AI safety and security research directions and to produce a data-driven ranking of their potential impact. These rankings may support evidence-based decisions about how to effectively deploy resources toward AI reliability and security research.

Problem

• Identify top AI reliability and security research priorities
• Quantify expert priorities for AI safety and security
• Rank research directions by potential impact for resource allocation

Innovation

• Surveyed 53 specialists across 105 AI areas
• Quantified expert priorities in AI safety
• Produced data-driven ranking for research impact