Harvey collects publicly available information, including court judgments, decisions, and public filings, and uses this information to develop, train, and improve its AI platform. This category of information is listed as a data source across multiple processing purposes throughout the policy.
This analysis describes what Harvey AI's agreement states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability and practical outcomes may vary by jurisdiction, enforcement context, and individual circumstances. Read our methodology
Publicly available information about individuals or entities may be incorporated into Harvey AI's training data without those individuals being Harvey AI customers or having any direct relationship with Harvey AI.
Interpretive note: The excerpt states two independent purposes for collecting publicly available information: (1) to offer and provide Services, and (2) to develop, train, and improve the AI platform. The canonical claim states the primary AI training purpose; the service-offering purpose is recorded in omitted_material.
Under this provision, publicly available information including judgments, decisions, and public filings may be collected and used to train Harvey's AI models. The agreement states that Harvey does not attempt to reidentify de-identified or aggregated data unless required by law, and a linked resource describes the sources and privacy impact minimization steps taken in connection with AI training data.
How other platforms handle this
Find and prevent fraud; and Block and remove unsafe or fraudulent users from the Lyft Platform.
any personal information that you provide through the Walmart Sites will be used by Walmart in accordance with Walmart's Privacy Policy
Where the law allows us to, we may use the content you and other users have posted for training or to help us to improve the way we filter content on our platform.
Monitoring
Harvey AI has changed this document before.
Receive same-day alerts, structured change summaries, and monitoring for up to 25 platforms.
"We collect publicly available information about customers and prospects to help offer and provide our Services. We use publicly available information (for example, judgments, decisions, public filings, etc.,) to develop, train, and improve our AI platform ("Publicly Available Information").— Excerpt from Harvey AI's Harvey AI Privacy Policy
(1) REGULATORY LANDSCAPE: This provision may require evaluation under GDPR's data minimization and purpose limitation principles (Articles 5 and 6) in the context of using publicly available personal data for AI training, particularly where individuals named in court documents or public filings retain data protection rights. The EU AI Act's provisions on training data transparency and documentation may also be relevant for AI systems deployed in professional legal and financial contexts. The FTC has issued guidance on AI training data practices under its unfair and deceptive practices authority. (2) GOVERNANCE EXPOSURE: Medium. The use of publicly available legal documents for AI training raises questions about whether individuals named in those documents were informed of or consented to this secondary use, and whether legitimate interests or other GDPR legal bases are sufficient to support this processing. The policy's reference to a linked page on training data sources and privacy minimization steps suggests awareness of this exposure but the sufficiency of those measures is not determinable from the policy alone. (3) JURISDICTION FLAGS: EU and UK data subjects named in publicly available documents may retain GDPR and UK GDPR rights even if the information is publicly accessible, as public availability does not automatically extinguish data protection rights. The right to object under GDPR Article 21 may apply where legitimate interests are the legal basis for training data processing. Some EU member states may have additional restrictions on processing publicly available personal data for AI development. (4) CONTRACT AND VENDOR IMPLICATIONS: Enterprise customers in legal and financial sectors should assess whether use of publicly available client-related documents in AI training is consistent with their own professional obligations and client confidentiality requirements. Law firms or financial institutions deploying Harvey should evaluate whether publicly available documents containing client information are included in training data and whether this creates professional responsibility or confidentiality considerations. (5) COMPLIANCE CONSIDERATIONS: Compliance teams should review the linked training data source page to assess what categories of public information are included and what privacy impact minimization measures are described. Organizations with GDPR obligations should evaluate whether the legitimate interests basis for this processing is adequately documented and whether data subject objection mechanisms are operationally supported for this processing purpose.
Full compliance analysis
Regulatory citations, enforcement risk, and due diligence action items.
Free: track 3 platforms + weekly digest. Monitor: 25 platforms + same-day alerts. No credit card required.
We read the privacy policies and terms of service of 38 AI platforms. Here is what they say about training, retention, arbitration, and liability.
Compliance Governance Intelligence
Need to monitor specific governance provisions?
Compliance includes provision-level monitoring, governance timelines, regulatory mapping, and audit-ready analysis.
Built from archived source documents, structured governance mappings, and historical version tracking.
Publicly available information about individuals or entities may be incorporated into Harvey AI's training data without those individuals being Harvey AI customers or having any direct relationship with Harvey AI.
Under this provision, publicly available information including judgments, decisions, and public filings may be collected and used to train Harvey's AI models. The agreement states that Harvey does not attempt to reidentify de-identified or aggregated data unless required by law, and a linked resource describes the sources and privacy impact minimization steps taken in connection with AI training data.
ConductAtlas has identified this type of provision across 280 platforms. See the full comparison.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Harvey AI.