Home › Best AI by category › Citation checking › Report
As of Oct 26, 2026, aiforlawyers.app counts 290 lawyer voices on citation checking; 246 voices name one of 12 tools, 4 of which have 10+ voices and are rated. Highest floor of the 90% band: Claude (net +16, band -12 to +41, 31 voices, Low confidence).
Download the PDF (cutoff Oct 26, 2026)Data (JSON)Ranking page
Cite this
aiforlawyers.app (AI Vortex), Citation checking report, cutoff Oct 26, 2026, https://aiforlawyers.app/report/citation-checking, licensed CC BY 4.0.
For one figure: “As of Oct 26, 2026, aiforlawyers.app counts 290 lawyer voices on citation checking; 246 voices name one of 12 tools, 4 of which have 10+ voices and are rated. Highest floor of the 90% band: Claude (net +16, band -12 to +41, 31 voices, Low confidence).”
12 tools named in 290 voices on citation checking. Rated tools first, ordered by the floor of the 90% band; then the tools with few voices, with their count and no band. Net = like minus dislike per 100 voices.
| # | Tool | Status | Voices | Net | 90% band | Confidence | Top complaint (voices) |
|---|---|---|---|---|---|---|---|
| 1 | Claude | Rated | 31 | +16 | −12 to +41 | Low | Hallucination (10) |
| 2 | Lexis+ AI | Rated | 17 | −47 | −71 to −9 | Low | Hallucination (6) |
| 3 |  ChatGPT | Rated | 157 | −69 | −76 to −59 | Medium | Hallucination (111) |
| 4 | Westlaw AI | Rated | 11 | −45 | −77 to +3 | Low | Hallucination (5) |
| – | Gemini | Few voices | 9 | – | no band: 9 of 10 voices needed | not rated | Hallucination (7) |
| – | CoCounsel | Few voices | 7 | – | no band: 7 of 10 voices needed | not rated | Hallucination (2) |
| – |  Harvey | Few voices | 6 | – | no band: 6 of 10 voices needed | not rated | Hallucination (4) |
| – | Perplexity | Few voices | 3 | – | no band: 3 of 10 voices needed | not rated | Hallucination (1) |
| – |  Copilot | Few voices | 2 | – | no band: 2 of 10 voices needed | not rated | Hallucination (1) |
| – |  Legora | Few voices | 1 | – | no band: 1 of 10 voices needed | not rated | none on record |
| – |  | Few voices | 1 | – | no band: 1 of 10 voices needed | not rated | none on record |
| – |  vLex Vincent | Few voices | 1 | – | no band: 1 of 10 voices needed | not rated | Hallucination (1) |
The most named complaints per tool, with a verbatim quote each. Counts are voices, not opinions of the firm or of us.
15 voices of 31 criticize it or are mixed about it with a named complaint: hallucination (10), accuracy (2), trust, liability (2), price (1).
“I've built in every conceivable safeguard that I can, and Claude's model just *consistently* underwhelm.”
“Lawyers representing #AI firm Anthropic in a lawsuit inadvertently put citation errors in their brief after using Anthropic's own tool, Claude, to format the material.”
12 voices of 17 criticize it or are mixed about it with a named complaint: hallucination (6), trust, liability (4), price (1), integration (1).
“When I tried to Sheperdize the cite, it became apparent this little nugget was cut from whole cloth.”
“What I find are real cases but misquoted and irrelevant. Protege is very expensive and a huge rip off.”
ChatGPTRated · 157 voices132 voices of 157 criticize it or are mixed about it with a named complaint: hallucination (111), accuracy (10), trust, liability (10), privacy (1).
“Duo from London law firm self-report to the SRA after filing fake quotes generated by ChatGPT and 'infected' with AI”
“ChatGPT isn’t capable of that kind of research yet.”
8 voices of 11 criticize it or are mixed about it with a named complaint: hallucination (5), accuracy (3).
“In the one case the client incorporated the changes in redline and mentioned that they were from Westlaw AI so I checked the citations and found a couple were inaccurate or misleading and (gently) told her that.”
“Running a brief through Westlaw cite check is insufficient IMO. It is just another AI platform.”
7 voices of 9 criticize it or are mixed about it with a named complaint: hallucination (7).
“Richard Ervin, the Attorney General of Florida, is supposed to have died in 1962, collapsing while arguing the case of Gideon v. Wainwright in front of SCOTUS. Fun fact, Ervin was AG of Florida until 1964, died in 2004.”
4 voices of 7 criticize it or are mixed about it with a named complaint: hallucination (2), accuracy (2).
“No cocounsel one-shots it that's why it didn't catch the hallucination.”
“More than once Westlaw CoCounsel has claimed a case held a proposition that one side merely asserted in the appeal, even though the judges didn’t reach that same conclusion, or may not have reached that issue at all.”
HarveyFew voices · 6 of 105 voices of 6 criticize it or are mixed about it with a named complaint: hallucination (4), trust, liability (1).
“All Harvey would need to do is query LexisNexis' system to see if those citations' exist in the database, and if they don't, refresh the generator.”
2 voices of 3 criticize it or are mixed about it with a named complaint: hallucination (1), accuracy (1).
“It found a problem in the citation but then hallucinated a different citation to use that does not match at all.”
“Even Perplexity that gives links to the sources, still gives citation wrong when I compare with the relating Rule.”
CopilotFew voices · 2 of 101 voice of 2 criticize it or are mixed about it with a named complaint: hallucination (1).
No quote on record that passes our quote rules (verbatim, a statement, no question).
LegoraFew voices · 1 of 10No named complaint on record among its 1 voice.
Few voices · 1 of 10No named complaint on record among its 1 voice.
vLex VincentFew voices · 1 of 101 voice of 1 criticize it or are mixed about it with a named complaint: hallucination (1).
No quote on record that passes our quote rules (verbatim, a statement, no question).
Public tests that include citation checking work, with their own publisher, date and caveat. They measure question sets, not lawyers’ experience, and are never added to the ranking above. Others are on the arena.
hallucination study · Stanford RegLab and HAI
published 2024-05-30 (arXiv v1; HAI news post dated 2024-05-23) · read 2026-10-22
The preregistered study found Lexis+ AI and Thomson Reuters' Westlaw AI-Assisted Research and Ask Practical Law AI hallucinated 17% to 33% of the time per the arXiv abstract, fewer than GPT-4. The HAI post puts Lexis+ AI and Ask Practical Law AI at more than 17% and Westlaw AI-Assisted Research at more than 34%.
Caveat. Tested 2024 versions; Westlaw range differs between abstract (up to 33%) and HAI post (above 34%); it covered only LexisNexis and Thomson Reuters products plus GPT-4 (no Harvey, Legora or Spellbook), and it ranks nothing overall.
Names: Lexis+ AI, Westlaw AI-Assisted Research, Ask Practical Law AI, GPT-4.
Related on AI Vortex
Use these numbers: the counts, rates and bands on this report are free to reuse under CC BY 4.0. Credit “aiforlawyers.app (AI Vortex)” with the date and a link to the page. Quotes belong to the people who wrote them: quote them with the link to the original post. Machine-readable: index.json. How to cite.
