Legal AI⚡ Limited Risk4/10

Harvey AI

AI assistant for law firms and legal teams for document review, contract drafting, legal research, and due diligence. Used internally by legal professionals with human review of all outputs.

⚡ EU AI Act Risk Classification

Indicative classification: Limited Risk

Legal reference: Article 50 — if client-facing; minimal/limited risk for internal legal professional use

Important: This is an indicative assessment based on typical deployment patterns. Your actual risk classification depends on your specific use case, configuration, and context. Register in Aeqora to get a precise classification using the EU AI Act decision tree.

Your obligations as a Deployer

→No mandatory EU AI Act obligations for internal legal team use with human review
→Article 50 chatbot disclosure if deployed to interact directly with clients
→Review if AI legal recommendations influence decisions on access to justice (context-specific Annex III risk)
→Data processing agreement review for EU client data

How this score was derived

Methodology →

Scores 4/10: Legal AI used by trained professionals with mandatory human review of all AI outputs. Internal legal workflow AI does not directly make consequential decisions about individuals — lawyers review and take responsibility for AI-assisted work product. Score of 4 reflects conditional Article 50 obligations for client-facing deployments and the context-specific risk that legal AI recommendations influencing access to justice may trigger Annex III review. No EU Declaration of Conformity published as of the index date.

Provider EU AI Act documentation

Harvey AI has not yet published dedicated EU AI Act compliance documentation. Contact them directly to request instructions of use and ask about their compliance roadmap.

What to do next

01

Register in Aeqora

Add Harvey AI to your AI system inventory and run the classification wizard for your specific deployment.

02

Request vendor docs

Ask Harvey AI for their EU AI Act instructions of use and technical documentation.

03

Generate compliance pack

Generate your Article 26 Deployer Compliance Evidence Pack once classification is complete.

Cite this assessment

Vendor? Request a correction →

Aeqora. "Harvey AI — EU AI Act Risk Assessment." Aeqora Vendor AI Risk Index. Verified 2026-06-14. https://aeqora.eu/vendors/harvey-ai

Scores are methodology-derived and citable under CC BY. View methodology →

Embed risk badge on your site

Add a live Aeqora risk badge to your vendor listing, procurement policy, or trust page. The badge stays current as our index is updated.

Limited Risk · EU AI Act← badge preview

<a href="https://aeqora.eu/vendors/harvey-ai" target="_blank" rel="noopener"> <img src="https://aeqora.eu/api/badge/harvey-ai" alt="Harvey AI EU AI Act risk classification" height="22"> </a>

Badge links to this page. See methodology for classification details.

Using Harvey AI?

Classify it precisely for your organization and generate your compliance documentation.

Start free — classify Harvey AI

← Back to all vendors