AI Policy

Harvey · agreement dated December 26, 2025 · Official agreement

Read the agreement

9 of 9 cards

Legal Rights & Agreement Rules · Context

The policy describes Harvey's current AI practices for app.harvey.ai

The AI Policy is Harvey's statement of its current practices and commitments on responsible AI use in connection with its service, defined as app.harvey.ai. It covers compliance with laws, responsible use, data sourcing, technical and organizational measures, transparency and risk management. The policy is dated December 26, 2025.

Supporting quotations

sets out Harvey’s approach to the development and deployment of artificial intelligence
(that is, app.harvey.ai)
This AI Policy serves as a statement of Harvey’s current practices and commitments in relation to the responsible use of AI.
Last updated: December 26, 2025

Qualifications & exceptions

  • Applies to the service defined as app.harvey.ai.

Not stated here

  • The policy does not say how or whether users are notified when it changes.
Also annotated in the same passage (1)

Legal Rights & Agreement Rules · Permission or license

Harvey may update the AI Policy from time to time

Harvey reserves the right to update or amend the AI Policy to reflect changes in technology, regulation or industry best practices. The policy states no notice period and no process for telling users about changes.

Supporting quotations

Harvey may update or amend this AI Policy from time to time to reflect changes in technology, regulation, or industry best practices.

Qualifications & exceptions

  • Reasons given: changes in technology, regulation or industry best practices.

Not stated here

  • No notice period or notification method is stated.
Also annotated in the same passage (1)

Legal Rights & Agreement Rules · Context

The policy is to be read with a separate AI acceptable use policy, if Harvey releases one

The AI Policy says it should be read alongside any AI acceptable use policy that Harvey releases from time to time, and points to Harvey's legal page. That policy's text is not part of this document, so what it requires is not covered here.

Supporting quotations

This AI Policy should be read alongside any AI acceptable use policy that Harvey releases from time to time

Qualifications & exceptions

  • Applies to an AI acceptable use policy only if and when Harvey releases one.

Not stated here

  • The text of any AI acceptable use policy is not supplied in this document.

Legal Rights & Agreement Rules · Company commitment

Harvey commits to comply with applicable AI laws, including the EU AI Act

Harvey says it will comply with applicable laws, regulations or directives relating to providing AI, naming the European Union's Artificial Intelligence Act (Regulation (EU) 2024/1689) as an example, and calls them together "AI Regulations". The commitment is limited to laws that apply to Harvey's provision of the Service.

Supporting quotations

To the extent applicable to Harvey’s provision of the Service, Harvey will comply with applicable laws, regulations, or directives in relation to the provision of AI, including the European Union’s Artificial Intelligence Act (Regulation (EU) 2024/1689)

Qualifications & exceptions

  • Only to the extent the laws are applicable to Harvey's provision of the Service.

Not stated here

  • The policy does not say which laws Harvey considers applicable, and the EU AI Act text is not supplied.

Security & Safeguards · Company commitment

Harvey says it will take steps on employee AI literacy and keep internal AI policies

Harvey says it will take steps to ensure a sufficient level of AI literacy among its employees who deal with operating and using AI in connection with the Service, in accordance with the AI Regulations. It also says it has implemented, and will maintain, internal policies and procedures for ethical and responsible AI use. The wording is about taking steps, not about guaranteeing a result.

Supporting quotations

Harvey will take steps to ensure a sufficient level of AI literacy of its employees dealing with the operation and use of AI in connection with the Service in accordance with the AI Regulations.
Harvey has implemented and will maintain internal policies and procedures for the ethical and responsible use of AI.

Qualifications & exceptions

  • Literacy steps cover employees dealing with the operation and use of AI in connection with the Service.

Not stated here

  • The policy does not describe what the literacy steps or the internal policies contain.

Content & IP Rights · Company commitment

Harvey says it will take steps to obtain rights to the data sets built into the Service

Harvey says it will take steps to obtain all necessary rights or licenses for the data sets incorporated into the Service, and will put procedures in place to monitor and verify compliance with third-party rights when collecting that data. The text speaks of data sets incorporated into the Service. It does not say whether this covers customer or user content.

Supporting quotations

Harvey will take steps to obtain all necessary rights and/or licenses required to use the data sets that are incorporated into the Service
implement procedures to monitor and verify compliance with third-party rights when collecting such data

Qualifications & exceptions

  • Covers data sets incorporated into the Service and collection of such data.

Not stated here

  • The policy does not say whether customer or user content counts as these data sets.
  • It does not say how the monitoring and verification procedures work.

Security & Safeguards · Company practice

Harvey lists example technical and organizational measures for its AI

Harvey says it implements and maintains a number of technical and organizational measures for developing and providing AI technologies in connection with the Service. The examples given are internal policies regulating AI use and development, security measures including those in its Security Addendum, testing to monitor and verify accuracy, reliability and suitability of the AI used for the Service, and change management processes aimed at continuing compliance with applicable laws and industry standards. The list is introduced as examples, so it is not a complete list.

Supporting quotations

Harvey implements and maintains a number of technical and organizational measures in relation to the development and provision of AI technologies in connection with the Service. Examples of such measures include:
internal policies to regulate the use and development of AI;
a number of security measures, including those described in our Security Addendum
testing to establish, monitor, and verify the level of accuracy and reliability and suitability of AI used for the purpose of the Service
change management processes to ensure continuing compliance with applicable laws and industry standards.

Qualifications & exceptions

  • The listed measures are examples, not an exhaustive list.
  • Security measures are those described in the Security Addendum, which is not supplied here.

Not stated here

  • The text of the Security Addendum is not supplied in this document.
  • The policy does not give testing results or accuracy levels.

Legal Rights & Agreement Rules · Company practice

Harvey has procedures to help customers tell people they are interacting with AI

Harvey says it has implemented procedures to assist you in ensuring that the natural persons you choose to have interact with the Service are informed that they are interacting with AI. The wording makes Harvey's role one of assisting you; it does not say Harvey itself gives that notice to those persons.

Supporting quotations

Harvey has implemented procedures to assist you in ensuring that those natural persons that you choose to interact with the Service are informed that they are interacting with AI.

Qualifications & exceptions

  • Applies to people whom the customer chooses to have interact with the Service.

Not stated here

  • The policy does not describe the procedures or what Harvey does to assist.

Legal Rights & Agreement Rules · Context

A previous version dated April 22, 2025 is linked

The page lists a previous version of the AI Policy dated April 22, 2025 and links to it. That earlier text is not part of this document, so this analysis does not say how it differs from the December 26, 2025 version.

Supporting quotations

April 22, 2025

Not stated here

  • The April 22, 2025 version is not supplied and is not compared.