Security controls for Generative AI

Vertex AI implements Google Cloud security controls to help secure your models and training data.

The following table lists the security controls available for Generative AI features:

	Data residency (at-rest)^1	Customer-managed encryption key (CMEK)	VPC Service Controls (VPC-SC)	Access Transparency (AXT)
PaLM 2 for Text
Online Prediction	✔	✔	✔	✔
Batch Prediction	✔	✔	✔	✔
Tuning	✔	✔^2	✔	✔
Evaluation	✔	✔	✔	✔
PaLM 2 for Chat
Online Prediction	✔	✔	✔	✔
Tuning	✔	✔^2	✔	✔
Evaluation
Codey for Code Generation
Online Prediction	✔	✔	✔	✔
Batch Prediction	✔	✔	✔	✔
Tuning	✔	✔^2	✔	✔
Evaluation
Embeddings for Text
Online Prediction	✔	✔	✔	✔
Batch Prediction	✔	✔	✔	✔
Tuning
Evaluation
Embeddings for Multimodal
Online Prediction	✔	✔	✔	✔
Imagen for Image Generation
Online Prediction	✔	✔	✔	✔
Tuning	✔	✔	✔	✔
Gemini 1.0 Pro
Online Prediction	✔	✔	✔	✔
Gemini 1.0 Pro Vision
Online Prediction	✔	✔	✔	✔

¹See Generative AI on Vertex AI locations for regions where the Generative AI on Vertex AI meets at-rest commitments.
²GPU only.

* Learn more about data residency by expanding General Service Terms and reading 1. Data Location.

Learn more

• Learn more about Vertex AI security controls.