HiveActs LogoHiveActs Back to site

AI Disclaimer

Last updated: 2026-05-28

HiveActs uses large language models (LLMs) supplied by Anthropic, PBC (the Claude family of models) to power contract analysis, clause-by-clause review, drafting from briefs, chat-based rewrites, and other features marked with the "AI" label inside the product. This page explains how that works and what its outputs are not.

The AI is not your lawyer

Nothing produced by the AI inside HiveActs is legal advice. It is a draft, a suggestion, a starting point. Anthropic Claude is a general-purpose language model. It can be wrong about the law in your jurisdiction, miss material context that wasn't in the prompt, confidently produce text that contradicts your company's playbook, or generate clauses that look right but reference rules that don't exist (so-called "hallucinations").

Every meaningful contractual decision must be reviewed by a qualified human lawyer admitted in the relevant jurisdiction before signature, before sending to a counterparty, or before relying on it in any regulated process. The Service does not establish an attorney-client relationship with you. We are not your lawyers.

What the AI is good at

  • Surfacing playbook deviations in a long contract you don't want to re-read line by line.
  • Generating a first-draft skeleton you can edit, instead of starting from a blank page.
  • Restating dense legal prose in plainer language.
  • Comparing wording against your own ideal/fallback clauses.
  • Answering structured questions about the document in front of it.

What the AI is bad at (be careful)

  • Knowing the current state of the law in your country, especially recent legislative changes.
  • Cross-jurisdictional comparisons. Numbers, citations, and case names can be fabricated.
  • Handling regulated industries (financial, medical, securities) where specific statutory wording matters.
  • Picking up issues hiding in attachments, exhibits, or references not present in the document text.
  • Tax, employment, antitrust, and immigration questions — these need a specialist.

How your data is used by the AI

  • When you trigger an analysis, chat message, or generation, we send the relevant document excerpts plus your playbook context to the Anthropic API.
  • Per Anthropic's API terms, this data is not used to train Anthropic's general-purpose models. Anthropic may retain prompts and outputs for up to 30 days for trust-and-safety purposes.
  • We don't store the prompt or AI response in our own database. We do store the resulting contract version, analysis items, or chat messages that your tenant explicitly creates.
  • Detailed flow lives in the Privacy Policy.

No guarantees

WE DO NOT WARRANT THAT AI OUTPUT IS ACCURATE, COMPLETE, UP TO DATE, OR FIT FOR ANY PARTICULAR LEGAL PURPOSE. You assume full responsibility for verifying any output before relying on it. See the Terms of Service for the full disclaimer and limitation of liability.

Reporting hallucinated output

If the AI generated something materially wrong — fabricated case law, a clause that misstates a regulation, contradictory legal advice — please send a copy to support@hiveacts.com. It helps us tune the playbook and prompts.

Models in use

Active model lineup is configurable per tenant. As of 2026-05-28 we route requests through Anthropic's Claude Haiku, Sonnet, and Opus families. We do not use any open-source local models, and we do not fine-tune any model on customer data.