Vanta’s AI principles: Building with trust and transparency

Today we announced Vanta AI, our suite of AI-powered tools to accelerate and simplify security and compliance workflows. With Vanta AI, tasks that were previously impossible to automate can now be performed reliably in minutes, enabling security and compliance teams to prove trust and manage risk more efficiently and confidently than ever before. 

‍

From the start, Vanta has been on a mission to secure the internet and protect consumer data. As we pioneer the use of AI in trust management, we are taking concerns around privacy, security, transparency, and accuracy very seriously.

‍

That’s why we’ve developed our AI principles to ensure that we build Vanta AI in a trustworthy and responsible manner — both for our team and the customers we serve today and in the future.

‍

Our AI principles outline how we plan to steward the safe and effective deployment of AI at Vanta. We’ve established an internal committee of leaders from our product, engineering, security, compliance, and legal teams responsible to ensure all new AI projects adhere to them, from conception to launch — and beyond. 

‍

Vanta’s AI principles are:

‍

1. Do no harm

All usage of AI at Vanta must first and foremost seek to do no harm to its customers or to Vanta. Reasonable attempts should be made to predict any and all potential cases of harm posed by a project.

‍

‍

2. Security and privacy by design

Every AI project and ongoing effort must incorporate security and privacy by design from day 0 and with every substantive change. Evidence must be shown.

‍

‍

3. Impact of incorrectness

Projects must commit to and have practical, achievable plans to assess the likelihoods and impacts of incorrectness and design human-in-the-loop review processes where necessary.

‍

‍

4. Explainability and transparency

Reasonable efforts are taken to ensure the explainability of results and provide transparency into the process by which they were derived.

‍

‍

5. Data control and risk

A clear understanding of the data being used by AI is established and guardrails are in place to control the scope of data access. A plan is established for the risks posed by such access as well as the resulting outputs.

‍

‍

These principles have been developed in line with the NIST AI Risk Management Framework and seek to move us iteratively toward further alignment with the framework and its intentions.

‍

We can’t wait for you to get started with Vanta AI. Learn more about how it works here. 

‍