Skip to main content
This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
MIT AI Risk Initiative
These visualizations are experimental and still being refined. Designs and data presentation may change.

Does Governance Reduce Harm?

Testing whether more governance correlates with fewer or less severe incidents. Toggle the Y-axis between incident count and expert BAU severity to see different perspectives on whether governance is associated with harm reduction.

X:
Y:
r = 0.32
Discrimination & Toxicity
Privacy & Security
Misinformation
Malicious Actors & Misuse
Human-Computer Interaction
Socioeconomic & Environmental
AI System Safety, Failures & Limitations

Key Takeaways

  • 1.Governance score vs. incident count: r = 0.32 (weak). More-governed subdomains tend to have more incidents (governance follows harm).
  • 2.Governance score vs. expert BAU severity: r = -0.00 (weak). Governance is not strongly linked to severity expectations.
  • 3.4.3 Fraud, scams, and targeted manipulation has the largest incident-to-governance ratio (464 incidents, score 251.0).
  • 4.6.5 Governance failure has the highest governance-to-incident ratio (score 398.5, 1 incidents).

Governance score weights enacted and proposed documents. Expert severity is BAU weighted mean (1–5).