All Collections
Usage instructions
'Ask AI' Module
Security and Privacy for 'Ask AI' Module
Security and Privacy for 'Ask AI' Module

FAQ's for 'Ask AI' Module Security and Privacy

John Paul Keeler avatar
Written by John Paul Keeler
Updated yesterday

Can IPRally employees see my Ask AI conversations?

No. As with other parts of the IPRally product, your data is encrypted and IPRally employees cannot see it in storage or transit.

What data does IPRally collect about my usage of the feature?

We collect statistics about usage of the feature in order to see how much it is used. We know that you asked a question but we do not know what the question was or what patent it was about. We also collect data on the lengths of the questions, lengths of the answers, and the sizes of the related patent documents.

Do you send my questions to a third party?

Yes. The current version of Ask AI Module is implemented using the Claude model from Anthropic. When you ask a question, we send the question to Anthropic’s servers.

We won’t send data to Anthropic or any other third party without your permission. In order to use the 'Ask AI' Module, your company’s IPRally admin user will need to accept the terms of the feature. After the admin user accepts the terms, they can enable the feature for your whole company.

What data do you send to Anthropic?

We send your question, the text of the patent, some other public information about the patent (the bibliographic data), and some custom instructions that help Claude to give better answers.

Some data we do not send: any identifiable information about you or your company (beyond what you write in the question), any of your tags, comments or other information you’ve entered in IPRally.

How do I know if I can trust Anthropic with our data?

We believe your data is secure with Anthropic. They adhere to the SOC 2 cybersecurity compliance framework. We encourage you to review their procedures at https://trust.anthropic.com/

What does Anthropic do with my data?

They feed in your question to a large language model to generate the answer.

They do not use the data to train their models or claim any ownership of the data.

They may retain your question for a set period of time for legal purposes, but do not use the question for anything beyond that.

Does Anthropic use my questions to train their AI?

No.

Can employees of IPRally or Anthropic see my questions?

No.

Why do IPRally's Generative AI Terms also mention OpenAI? I thought you used a model by Anthropic, not OpenAI?

We wanted to include OpenAI in the terms in case we decide to base features on their services later on. Currently we are using OpenAI for some internal features in development but have not released any such features publicly to our users.

You can review the procedures of OpenAI at https://trust.openai.com/

Did this answer your question?