Google To Develop Free Internal Video Blurring Tool Called Magritte

Andi C., Tech Times 22 December 2022, 06:12 pm

Google's latest internal video blurring tool, called Magritte is under development and will reportedly be free to use.

According to Engadget, two of Google's latest privacy-enhancing technologies (PETs), including one that blurs objects in videos, will be made available to everyone for free via open source.

Google reveals that the new capabilities are a part of its Protected Computing program, which aims to change how data is processed to secure its privacy and safety technically.

PETs are a category of technologies that will enable researchers, doctors, and other anyone with access to sensitive data to gain insights even without seeing the data itself, according to the White House.

However, the real strength of PETs lies in their ability to analyze data while keeping it "hidden" from researchers. PETs might open the door to new kinds of teamwork and new standards for the responsible handling of personal data.

Google's Internal 'Magritte': What We Know So Far

The first is an internal PET project is called Magritte, which is now out on Github. Magritte uses machine learning to detect objects and apply a blur as soon as they appear on the screen.

Along with the launch of Magrittle, a new version of the open-source tool FHE C++ Transpiler has also been unveiled. Researchers' interest in the technology has increased recently as a result of its potential to increase the security of enterprise applications.

Silicon Angle reports that it can cover up different symbols and privacy-sensitive details such as registration plates, tattoos, and more. Google stated that this code is convenient for video journalists who want to provide enhanced privacy safeguards or videographers who will need to save time from blurring items from a video. The open-source code can reportedly do detection across a video with high accuracy.

Google's Privacy Enhancing Technologies (PET)

Aside from Magritte, Google will also release another PET called "Fully Homomorphic Encryption (FHE) Transpiler," which allows developers to perform computations on encrypted data without accessing personally identifiable information.

Help Net Security reports that the term "FHE" refers to a set of encryption techniques that differ from conventional encryption techniques. They enable computation directly on encrypted data without requiring a secret key.

Google says it can help different industries, such as financial services, healthcare, and government, where a robust security guarantee around the processing of sensitive data is of the highest importance.

The development of PETs might make it easier for people to work together across boundaries, industries, and institutions to address common problems, advancing solutions in fields including pandemic preparedness, financial crime, human trafficking, and health care.

However, as of now, PETs have not yet been widely used due to a number of reasons, including the need for greater research and development, a lack of technical knowledge, perceived and potential hazards, financial costs, and the absence of solutions that might be applied universally.