Multi-Perspective Risk Assessment: Integrating Societal, Media, and Technical Signals for Holistic AI Governance

While technical audits focus on quantifiable risks (Casper et al., 2024; Ilori et al., 2024), and news media shape public perceptions of AI risks differently across contexts (Allaham et al., 2025), there’s little integration between these streams. This research would build a framework that systematically combines audit findings, news/media analysis (using NLP to assess framing and risk salience, as in Allaham et al.), and direct stakeholder input (citizens, affected groups, etc.). By triangulating these sources, regulators and organizations can identify risks that are technically present but societally invisible—or vice versa. The novelty lies in operationalizing “context-aware” governance: for example, flagging risks that technical audits miss but that are prominent in societal debate, or mediating policy interventions based on alignment/misalignment between technical and social risk signals. Such a framework could help bridge the “governance gap” by ensuring that risk assessments are not just technically rigorous, but also socially and politically legitimate.

References:

1. Advanced data analytics in internal audits: A conceptual framework for comprehensive risk assessment and fraud detection. Oluwatosin Ilori, Nelly Tochi Nwosu, Henry Nwapali Ndidi Naiho (2024). Finance & Accounting Research Journal.
2. Black-Box Access is Insufficient for Rigorous AI Audits. Stephen Casper, Carson Ezell, Charlotte Siegmann, Noam Kolt, Taylor Lynn Curtis, Ben Bucknall, Andreas A. Haupt, Kevin Wei, Jérémy Scheurer, Marius Hobbhahn, Lee Sharkey, Satyapriya Krishna, Marvin von Hagen, Silas Alberti, Alan Chan, Qinyi Sun, Michael Gerovitch, David Bau, Max Tegmark, David Krueger, Dylan Hadfield-Menell (2024). Conference on Fairness, Accountability and Transparency.
3. Informing AI Risk Assessment with News Media: Analyzing National and Political Variation in the Coverage of AI Risks. Mowafak Allaham, Kimon Kieslich, Nicholas Diakopoulos (2025). Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society.