Question
How does Zendesk handle data privacy for AI products? I need a copy of the Data Processing Agreement (DPA).
How confident are you at Zendesk that your LLM (large language model) tool can accurately identify health data shared by users? And what safeguards do you have in place to protect this sensitive information?
Answer
Zendesk follows design principles for AI products that prioritize privacy, security, and compliance by design. For more details, refer to the Data Processing Agreement.
Zendesk sets a confidence threshold to find data and shows how sure it is about each message to reduce errors. You can adjust this threshold to balance accuracy and responses.
It's important to note that Zendesk uses proprietary machine learning models that are not generative in nature. Because the outputs of these models are labels rather than passages of free text, images, or code, they do not create the risk of customer data being reproduced or shared in an improper context.
For more information, see the following articles:
- Zendesk AI Data Use Information
- Complying with Privacy and Data Protection Law in AI agents - Advanced (Ultimate)
- Complying with Privacy and Data Protection Law in Zendesk products
- Generative AI at Zendesk