Data Breach

Microsoft AI research division accidentally exposed 38TB of sensitive data

Microsoft AI researchers accidentally exposed 38TB of sensitive data via a public GitHub repository since July 2020.

Cybersecurity firm Wiz discovered that the Microsoft AI research division accidentally leaked 38TB of sensitive while publishing a bucket of open-source training data on GitHub.

The exposed data exposed a disk backup of two employees’ workstations containing secrets, private keys, passwords, and over 30,000 internal Microsoft Teams messages. 

“The researchers shared their files using an Azure feature called SAS tokens, which allows you to share data from Azure Storage accounts.” reads the report published by Wiz.”The access level can be limited to specific files only; however, in this case, the link was configured to share the entire storage account — including another 38TB of private files.”

Wiz Research Team discovered the repository while scanning the Internet for misconfigured storage containers exposing cloud-hosted data. The experts found a repository on GitHub under the Microsoft organization named robust-models-transfer.

The repository belongs to Microsoft’s AI research division, which used it to provide open-source code and AI models for image recognition. The Microsoft AI research team started publishing data in July 2020.

Microsoft used Azure SAS tokens to share data stored in Azure Storage accounts used by its research team.

The Azure Storage signed URL used to access the repository was mistakenly configured to grant permissions on the entire storage account, exposing private data. 

“However, this URL allowed access to more than just open-source models. It was configured to grant permissions on the entire storage account, exposing additional private data by mistake.” continues the company. “The simple step of sharing an AI dataset led to a major data leak, containing over 38TB of private data. The root cause was the usage of Account SAS tokens as the sharing mechanism. Due to a lack of monitoring and governance, SAS tokens pose a security risk, and their usage should be as limited as possible.”

Wiz pointed out that SAS tokens cannot be easily tracked because Microsoft does not provide a centralized way to manage them within the Azure portal.

Microsoft said that the data lead did not expose customer data.

“No customer data was exposed, and no other internal services were put at risk because of this issue. No customer action is required in response to this issue.” reads the post published by Microsoft.

Below is the timeline of this security incident:

  • Jul. 20, 2020 – SAS token first committed to GitHub; expiry set to Oct. 5, 2021
  • Oct. 6, 2021 – SAS token expiry updated to Oct. 6, 2051
  • Jun. 22, 2023 – Wiz Research finds and reports issue to MSRC
  • Jun. 24, 2023 – SAS token invalidated by Microsoft
  • Jul. 7, 2023 – SAS token replaced on GitHub
  • Aug. 16, 2023 – Microsoft completes internal investigation of potential impact
  • Sep. 18, 2023 – Public disclosure

Follow me on Twitter: @securityaffairs and Facebook and Mastodon

Pierluigi Paganini

(SecurityAffairs – hacking, Microsoft AI)

Pierluigi Paganini

Pierluigi Paganini is member of the ENISA (European Union Agency for Network and Information Security) Threat Landscape Stakeholder Group and Cyber G7 Group, he is also a Security Evangelist, Security Analyst and Freelance Writer. Editor-in-Chief at "Cyber Defense Magazine", Pierluigi is a cyber security expert with over 20 years experience in the field, he is Certified Ethical Hacker at EC Council in London. The passion for writing and a strong belief that security is founded on sharing and awareness led Pierluigi to find the security blog "Security Affairs" recently named a Top National Security Resource for US. Pierluigi is a member of the "The Hacker News" team and he is a writer for some major publications in the field such as Cyber War Zone, ICTTF, Infosec Island, Infosec Institute, The Hacker News Magazine and for many other Security magazines. Author of the Books "The Deep Dark Web" and “Digital Virtual Currency and Bitcoin”.

Recent Posts

Fintech firm Figure disclosed data breach after employee phishing attack

Fintech firm Figure confirmed a data breach after hackers used social engineering to trick an…

17 hours ago

U.S. CISA adds a flaw in BeyondTrust RS and PRA to its Known Exploited Vulnerabilities catalog

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) adds a flaw in BeyondTrust RS and…

18 hours ago

Suspected Russian hackers deploy CANFAIL malware against Ukraine

A new alleged Russia-linked APT group targeted Ukrainian defense, government, and energy groups, with CANFAIL…

23 hours ago

New threat actor UAT-9921 deploys VoidLink against enterprise sectors

A new threat actor, UAT-9921, uses the modular VoidLink framework to target technology and financial…

1 day ago

Attackers exploit BeyondTrust CVE-2026-1731 within hours of PoC release

Attackers quickly targeted BeyondTrust flaw CVE-2026-1731 after a PoC was released, enabling unauthenticated remote code…

2 days ago

Google: state-backed hackers exploit Gemini AI for cyber recon and attacks

Google says nation-state actors used Gemini AI for reconnaissance and attack support in cyber operations.…

2 days ago

This website uses cookies.