A new privacy tool from Google has been made publically accessible, the company has revealed.

In a post on the Google Developers blog, the company revealed that the Magritte tool would be released as the newest addition to Google’s Protected Computing initiative. This program is meant to radically alter “how, when, and where data is processed to technically ensure its privacy and safety,” according to Google.

This new application employs “high-accuracy” machine learning to detect identifying items like license plates and tattoos and then automatically blurs them away before uploading pictures to the open-source project repository Github.

According to Google, Magritte is the finest tool for helping photographers and journalists protect the privacy of those they film. It was emphasized that it requires little in the way of computing resources and that its high degree of accuracy makes it a trustworthy instrument for economizing one’s time.

In other news, the less catchy “Fully Homomorphic Encryption (FHE) Transplier,” which was first released last year and aims to let data scientists compute encrypted data without accessing personal information, has received new circuit optimizations to ensure a lower computational cost and increase its use cases.

These resources are only the most recent illustration of Google’s commitment to privacy-enhancing technology (PET) research and development, which became a priority for the United States Government’s Office of Science and Technology Policy in June 2022. (OSTP).

The idea was also included in a competition hosted by the governments of the United States and United Kingdom in November 2022, in which contestants were tasked with coming up with ways to train AI models without compromising individual privacy.

The Privacy on Beam privacy framework, for example, requires no prior knowledge of differential privacy, but in 2019 Google made their differential privacy library accessible on GitHub so that even non-experts could use it.