When Open-Sourced models in Azure AI Foundry model catalog, does Microsoft do any additional security checks?

Question

When Open-Sourced models in Azure AI Foundry model catalog, does Microsoft do any additional security checks?

Benny Ng 0 Microsoft Employee

I saw that the following Open-Sourced model did not get scanned.
Does this mean it is not safe to use?

When deploying such open source models, does Microsoft do any additional security checks?

Is the model downloaded at real-time from the Open-Sourced origin or there is a pre-downloaded model stored within Microsoft? if so, does Microsoft do any security checks to ensure it is safe before storing?

https://ai.azure.com/explore/models/sentence-transformers-all-minilm-l6-v2/version/3/registry/HuggingFace

kothapally Snigdha 3,020 Reputation points Microsoft External Staff Moderator

2025-03-21T00:34:08.4266667+00:00

Hi Benny Ng

When deploying models from the Azure AI Foundry model catalog, Microsoft does implement security measures, including scanning for vulnerabilities that could exfiltrate data. However, not all models available through the model catalog have been scanned. This means that while some models may have undergone security checks, others may not, and caution should be exercised when using models that have not been scanned.

Regarding the storage of models, Microsoft does not specify whether models are downloaded in real-time from their open-sourced origins or if they are pre-downloaded and stored. However, it is important to note that Microsoft has data, privacy, and security commitments that apply to Azure Machine Learning, which includes measures for protecting deployments.

https://learn.microsoft.com/en-us/azure/machine-learning/concept-data-privacy?view=azureml-api-2

I hope these helps you. Thank you!
kothapally Snigdha 3,020 Reputation points Microsoft External Staff Moderator

2025-03-24T20:19:08.1933333+00:00

Hi Benny Ng

We haven’t heard from you on the last response and was just checking back to see if you have a resolution yet. In case if you have any resolution, please do share that same with the community as it can be helpful to others. Otherwise, will respond with more details and we will try to help.

Your answer

kothapally Snigdha 3,020 Reputation points Microsoft External Staff Moderator

2025-03-21T00:34:08.4266667+00:00

Hi Benny Ng

When deploying models from the Azure AI Foundry model catalog, Microsoft does implement security measures, including scanning for vulnerabilities that could exfiltrate data. However, not all models available through the model catalog have been scanned. This means that while some models may have undergone security checks, others may not, and caution should be exercised when using models that have not been scanned.

Regarding the storage of models, Microsoft does not specify whether models are downloaded in real-time from their open-sourced origins or if they are pre-downloaded and stored. However, it is important to note that Microsoft has data, privacy, and security commitments that apply to Azure Machine Learning, which includes measures for protecting deployments.

https://learn.microsoft.com/en-us/azure/machine-learning/concept-data-privacy?view=azureml-api-2

I hope these helps you. Thank you!
kothapally Snigdha 3,020 Reputation points Microsoft External Staff Moderator

2025-03-24T20:19:08.1933333+00:00

Hi Benny Ng

We haven’t heard from you on the last response and was just checking back to see if you have a resolution yet. In case if you have any resolution, please do share that same with the community as it can be helpful to others. Otherwise, will respond with more details and we will try to help.

Share via

When Open-Sourced models in Azure AI Foundry model catalog, does Microsoft do any additional security checks?

Your answer