Legal AI
Paxton AI
Medium riskBudget-priced all-in-one AI legal assistant (research, drafting, document review) marketed to solo practitioners and small firms.
Is Paxton AI safe for confidential data?
Paxton is unusually transparent for a budget tool: it published benchmark results against the Stanford Legal Hallucination Benchmark (roughly 94% average accuracy on its selected task sample, with a confidence-indicator feature), and its security page claims SOC 2 attestation, ISO 27001 and HIPAA-aligned controls, encryption, and no training on user data. The risks are the mirror image of the pitch. A 94% score means roughly one answer in twenty is wrong, and Paxton's price point puts it in the hands of solo lawyers with no research staff to catch that error before it reaches a filing — the exact profile overrepresented in the 1,600+ court decisions in Charlotin's AI-hallucination database. The benchmark sample was also selected by Paxton itself, and its compliance claims rest on the vendor's own page rather than widely available third-party reports; ask for the SOC 2 report before uploading client files.
Risk by plan
The same product often carries very different terms depending on the tier — consumer plans are where the exposure concentrates.
Low per-seat pricing means adoption is often an individual lawyer's decision on a personal card, without any firm-level review of the terms.
Data handling
Training on inputs
Paxton states user data, including uploaded documents and workflows, is never used to train models and is not shared beyond the customer's organization; terms appear uniform across plans.
Retention
Uploaded documents are stored encrypted in the customer's workspace; Paxton does not publish a detailed retention/deletion schedule — confirm deletion terms before uploading client files.
Residency
Not published; assume U.S. processing unless your agreement states otherwise. Canadian solos subject to PIPEDA should confirm cross-border handling.
Compliance
- SOC 2Yes
- GDPR / DPANot verified
- HIPAA BAANot verified
Certifications typically apply to specific tiers and contracts — confirm scope in writing before relying on them.
New to these frameworks? See our plain-language guides to SOC 2 and the other AI compliance standards.
Enterprise controls
- Encryption in transit and at rest
- SOC 2 attestation (announced by vendor; request the report)
- Confidence indicator on research answers
- Organizational workspaces and admin features
Frequently asked questions
How accurate is Paxton AI?
Paxton self-reported roughly 94% average accuracy on about 1,600 tasks it selected from the Stanford Legal Hallucination Benchmark, plus 94% on the CaseHold benchmark for its citator. That is a genuinely strong, unusually transparent result — and it still implies an error roughly every twentieth answer. The vendor chose the task sample, so treat the number as indicative, and verify every citation independently before filing.
How does Paxton AI handle data security?
Paxton's security page describes SOC 2, ISO 27001 and HIPAA-aligned controls, encryption in transit and at rest, and a commitment that user data is never used for model training. These claims come from the vendor's own site; for client-confidential work, request the actual SOC 2 report and a signed agreement restating the no-training term rather than relying on the marketing page.
Is a cheap AI assistant like Paxton a shadow-AI risk?
Often, yes — not because of the tool but because of how it arrives: an individual lawyer subscribes on a personal card, uploads client documents, and the firm never sees the terms. If Paxton (or any peer tool) is in use, move it onto a firm-level agreement, and assume the same lawyers are also using free consumer chatbots that deserve the same scrutiny.
Policy changelog
- Initial entry published from Paxton's published security documentation and cited coverage.
Sources
- Paxton security and compliance page
- Paxton benchmark announcement (94% on Stanford benchmark)
- LawSites coverage of Paxton's benchmark release
This profile summarizes the vendor's published policies as of the verification date. It is not legal advice.
Paxton AI is probably already in your organization.
Sanitized AI shows you who is using it and redacts sensitive data from prompts before it leaves your control.
More AI tool profiles
AI assistant built into Clio Manage, the dominant solo and small-firm practice management platform, answering questions and drafting from the firm's own matter data.
Thomson Reuters' professional AI assistant, delivered inside Westlaw Precision for legal work and as CoCounsel Tax, Audit & Accounting for advisory practices.
Legal AI platform for law firms and in-house teams: research, drafting, document review, and workflow automation built on foundation models.
LexisNexis' generative AI research and drafting assistant, now sold as Lexis+ with Protégé, grounded in the Lexis primary-law and Shepard's citation database.