The Google team announced that two of its internal privacy tools have become open source.
Tools that ensure the integrity of private data taking into account different contexts. We tell you what it is about.
An AI tool that blurs faces, license plates and more
Google has released two of its internal privacy tools as open source. Two initiatives that help process user data while staying within the limits of privacy.
For example, one of the tools that will have an open source version is Magritte. A tool that can automatically detect when an object on video contains sensitive data, such as a license plate.
When this happens, this tool blurs objects as they appear on the screen, thanks to the work of machine learning, as Google explains:
[…] We are pleased to announce an open source version of an internal project, Magritte, which uses advances in Machine Learning (ML) to detect objects that use few computational resources and applies a blur to those objects automatically as soon as they appear on the screen. The tool can blur arbitrary objects such as license plates and more.
As the Google team mentions, this tool could be especially useful for video journalists, since it will make it easier for them to protect privacy. They won’t have to waste time blurring objects or data in the video, as the algorithm can do that job automatically.
On the other hand, the second tool that goes to open source is what they call Fully Homomorphic Encryption (FHE) Transpiler. This tool has technology that allows you to work with encrypted data without having to access personally identifiable information.
A dynamic that could be useful in different sectors, such as health care or financial services, since they have to work with and process large amounts of sensitive data.
