| citations | public bibliographic records (CrossRef, OpenAlex, Open Library, DOI registries) | yes |
|---|---|---|
| quotes | a source text supplied by the user | no |
| figures | materials supplied by the user | no |
| gate | a policy about the quality of sources | no |
| grounding | a list of claims the user approved | no |
| values | a statistical source's own data (World Bank, IMF, Eurostat, Our World in Data, any statistics service the user describes, or seeded figures) | yes |
| links | the live web (server-side only) | yes |
| support (0.4.0) | a source text, via a caller-supplied judge whose evidence is verified | no (the judge may) |
| retractions (0.4.0) | Crossref's notices, including Retraction Watch | yes |
| legal (0.4.0) | CourtListener's database of US decisions (server-side only) | yes |
AI & evaluation · Updated 2026-11
The checks in Assay and what each compares text with
Table 1 in “Assay: Checking AI-Written Text Against Evidence, One Question at a Time”: 10 rows, 3 columns.
Cite this chart
Extrafemi Research (2026). The checks in Assay and what each compares text with [Chart]. Data from Assay: Checking AI-Written Text Against Evidence, One Question at a Time. https://research.extrafemi.com/data/assay--data-1