Frequently Asked Questions: Vanta AI

  • Updated

How do I get access to Vanta AI?

  • Vanta AI is currently in beta and available to paying Vanta customers via the products they have signed up for. For example, Vanta AI for reviewing vendor security documentation is available to customers who have purchased our Vendor Risk Management offering. 

Which technologies does Vanta AI use?

  • Vanta AI uses a combination of self-hosted Large Language Models (LLMs) as well as those available via API from OpenAI. 

How does Vanta AI share my data with AI models?

  • Vanta uses secure APIs to communicate with OpenAI. Vanta has an executed DPA with OpenAI to ensure that customer data is not stored or used for training. 

Vanta has architected our AI usage to ensure that AI features do not allow for any inappropriate data access across users or customers.

 

What security measures do you have in place for Vanta AI?

  • Internally hosted LLMs are hosted and deployed in Vanta’s cloud environment using industry best practices. Models hosted by vendors such as OpenAI are accessed via secure APIs. Please refer to Vanta’s security page and our Trust Report for more information.

How reliable is Vanta AI?

  • We strive to ensure quality and accuracy in Vanta AI’s performance through regular automated and human checks. However, as a new technology in open beta, occasional inaccuracies or availability issues might arise. If you find an issue, please report it via in-product feedback or support.