Got 2 minutes? We'd love your feedback to help us improve our documentation.

AI Product Addendum

Prev Next

This AI Product Addendum (the "Addendum"), once agreed with a client (the "Client"), will form part of the custom Subscription Services Agreement (the "Agreement") under which Natterbox Limited ("Natterbox") provides its subscription-based cloud and network services to the Client.

    1. Capitalised words and phrases defined elsewhere in the Agreement and used but not defined in this Addendum shall have the same meanings in this Addendum as in the part of the Agreement in which they are defined.

      1. Acceptable Use Restrictions: The Client must not use the AI Services, or any outputs generated by them, to: (a) develop, train, or fine-tune competing foundation models or artificial intelligence systems; (b) reverse engineer, decompile, or extract the underlying AI models; or (c) generate content that is illegal, harmful, abusive, or otherwise prohibited by the respective third-party acceptable use policies.

      2. Geographic Limits: The Client's access to and use of specific AI models (such as those provided via AWS Bedrock) is limited to geographic regions and countries supported by the respective third-party providers.

      3. Independent Verification: The Client acknowledges that, in accordance with Anthropic’s commercial terms, AI-generated outputs may be inaccurate, incomplete, or misleading, and must be independently verified by the Client before any reliance or use in a live environment.

      4. Third-Party Warranties Disclaimer: In addition to the limitations in Section 3, Natterbox expressly disclaims on behalf of its third-party providers (including AWS and Anthropic) any implied warranties regarding the accuracy, completeness, or fitness for a particular purpose of AI-generated content.

      5. Third-Party Data Privacy: In addition to Section 2, Natterbox confirms that Client Data processed through AWS Bedrock is isolated to Natterbox’s environment and is strictly not utilised by AWS or Anthropic to train, improve, or fine-tune their own foundational models.