These visualizations are experimental and still being refined. Designs and data presentation may change.
Governance Lifecycle Coverage
How governance attention is distributed across the 6 AI lifecycle stages — from planning and design through deployment and monitoring. Reveals which stages attract the most regulatory focus and where coverage gaps may exist.
Unspecified
Source: MIT AI Risk Repository
Key Takeaways
1.Operate & Monitor (843 docs) and Deploy (823) are the most-regulated stages — roughly -2% more governance attention than the three upstream stages (Plan & Design, Verify & Validate, Build & Use).
2.Collect & Process has the thinnest coverage at 441 documents — about half that of Operate & Monitor — despite its importance to downstream risk.
3.Hard Law accounts for 0–0% of documents in every stage, meaning binding regulation spans the entire lifecycle uniformly — no phase is dominated by soft law.
4.285 documents (31%) apply to all six lifecycle stages simultaneously, and governance documents touch 4.1 stages on average — most regulation is broad-spectrum rather than stage-specific.
