Up to 38TB! Microsoft AI team leaks massive data, including more than 30,000 internal messages

[CNMO News] Recently, according to CNMO, a cloud security startup called Wiz Research released a research report, pointing out that when rewriting content, the language needs to be rewritten into Chinese, and artificial intelligence (AI) that does not require the original sentence to appear ) The research team’s large cache of private data on GitHub was accidentally leaked. The incident, triggered by a misconfigured SAS token, exposed 38 terabytes of additional internal data, which contained backups of two employees' personal computer disks when content was rewritten so the language needed to be changed to Chinese so that the original sentences did not appear.

When rewriting the content, the language needs to be rewritten into Chinese, and the original sentence does not need to appear

It is understood that when rewriting the content, the language needs to be rewritten to Chinese and the original sentence does not need to appear. The AI ​​research team originally released open source training data on GitHub, but due to the misconfiguration of the SAS token, the data was accidentally exposed. The exposed data included secrets, private keys, passwords and more than 30,000 internal Microsoft Teams messages.

In this regard, when rewriting the content, the language needs to be changed to Chinese, and the original sentence does not need to appear in response, indicating that the SAS token has been revoked and an internal investigation of the potential impact has been completed. They emphasized that no customer data was exposed as a result, and there was no need for other rewrites to change the language to Chinese. There was no need for the original sentence service to be put at risk due to the issue.

However, cloud security startup Wiz Research believes that this incident reminds engineers that they must fully consider the required security measures when organizing artificial intelligence research. Even a technology giant that needs to rewrite the content so that the language needs to be changed to Chinese and the original sentence does not need to appear can lead to such a large-scale data leak because of a small configuration error.

This incident not only had an impact on the need to rewrite the content into Chinese without the original sentence itself, but also served as a warning to the entire technology industry. Driven by emerging technologies such as artificial intelligence and cloud computing, how to protect privacy and data security has become a problem that every technology company must face.

The above is the detailed content of Up to 38TB! Microsoft AI team leaks massive data, including more than 30,000 internal messages. For more information, please follow other related articles on the PHP Chinese website!