Microsoft AI researchers accidentally exposed terabytes of internal sensitive data

In September 2023, the giant Microsoft came under the spotlight for the umpteenth time, and not for good reasons. Inside the company, some AI researchers inadvertently posted tens of terabytes of sensitive data online, including passwords, private keys and more than 30,000 internal Microsoft Teams messages. While uploading open source training data buckets to GitHub, the perpetrators shared the mentioned private data.

The situation is not the best: 38 terabytes of information, including personal computer backups of two Microsoft employees, were made public. To make matters worse, it was discovered that the URL containing the mentioned data had existed since 2020 and had been misconfigured - anyone in possession of the latter had the ability to delete, replace and inject malicious content.

AI is undoubtedly of considerable importance within this industry, which is why it should be crucial to ensure a certain level of security when handling the private data of billions of users.