Methodology
How ranking works on AI Atlas right now — simple, transparent, and free.
Current ranking (MVP)
Problems are ranked primarily by community upvotes. Each signed-in member can upvote a published problem once. Rankings also surface newest and emerging problems (recently published) and problems with stronger evidence signals.
Most upvoted → upvote_count (desc), then published_at Newest → published_at (desc) Emerging → published within ~45 days, sorted by upvotes Most evidence → evidence_quality_score (desc)
What counts as contribution
- Posting a structured problem (AI-assisted drafting)
- Upvoting published problems
- Commenting with context or caveats
- Adding evidence links (reports, data, articles)
What comes later
Multi-dimensional scoring (importance, urgency, reach, AI suitability, feasibility, and willingness to adopt) with confidence and country balancing remains in the schema for a future release. It does not drive the product UI in this MVP.
See live rankings and explore.