Microsoft is testing the detection and removal of sensitive info from documents via Azure AI

Microsoft continues to add AI-based features in its many cloud-based services, including Azure. Today, the company revealed a new feature in Azure AI that it is testing with a limited number of businesses.

In a blog post, Microsoft stated that in the past, businesses that wanted to use AI on documents for things like summaries or removing content that has been labeled for redaction required them to crack open these documents as data. The text is then extracted from the data, and any text that needs to be redacted is removed. Finally, the document is recreated in its original format with the redacted or summarized info.

Microsoft says this method has become "time-consuming, expensive, and inconvenient." Today, the company is letting businesses apply for access to their public preview of a new method that allows businesses to simply upload the native document to Azure AI, where it will then be able to summarize or redact content with its REST API for PII. It can then bring the document back to its file format.

Healthcare professionals could use this to upload patients' records and redact or summarize their info before they are stored and archived. Insurance companies could do the same for a customer's policy information. Financial institutions could use it to remove tentative info for customers, like their account info or their loan application.

All of these scenarios with this feature will help businesses store and secure sensitive customer information and boost the confidence of their clients that their private info will remain private.

Businesses can also customize this new REST API for PII service with different parameters depending on their need. These may include redacting specific categories in documents, particular characters, or if the redacted text should be shown with characters or respective entity names.

There's no word yet on when this new Azure AU feature will become generally available.