Skip to content

Deepseek with Amazon Bedrock

0

Recent news regarding deepseek point to many governments around the world blocking its usage. With that in mind, what are the security technical controls in place in this deployment format that prevent customer data from being sent back to servers in control of the model owner and/or prevent data from being utilized to train their models?

1 Answer
1

The Bedrock FAQ answers this question and it applies to all models hosted in Amazon Bedrock. https://aws.amazon.com/bedrock/faqs/#:~:text=With%20Amazon%20Bedrock%2C%20your%20content,data%20using%20your%20own%20keys.

Security:

Q: Are user inputs and model outputs made available to third-party model providers?
A: No. Users' inputs and model outputs are not shared with any model providers.


Q: Will AWS and third-party model providers use customer inputs to or outputs from Amazon Bedrock to train Amazon Nova, Amazon Titan or any third-party models?
A: No, AWS and the third-party model providers will not use any inputs to or outputs from Amazon Bedrock to train Amazon Nova, Amazon Titan, or any third-party models.


Additionally, see this AWS Machine Learning Blog post specifically addressing Bedrock security with the Deepseek models.
Protect your DeepSeek model deployments with Amazon Bedrock Guardrails

AWS
EXPERT

answered 2 years ago

EXPERT

reviewed 2 years ago

  • I'm more interested in understanding the technical controls that prevent data movements that could either be used for sharing data or training models. It's good to read that "...AWS and the third-party model providers will not use any inputs to or outputs..." - however, it would be more supportive to understand how these preventions are technically enforced.

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.