Google has announced that it will be releasing two of its latest privacy-enhancing technologies (PETs), including a tool that blurs objects in a video, as open source for anyone to use for free. These tools are part of Google’s Protected Computing initiative, which aims to transform “how, when and where data is processed to technically ensure its privacy and safety,” according to the company.
The first of these technologies is called Magritte, which is now available on GitHub. Magritte uses machine learning to detect objects in a video and apply a blur as soon as they appear on the screen. It can be used to obscure objects such as license plates and tattoos. According to Google, this tool is particularly useful for video journalists who want to provide additional privacy assurances. By using this open-source code, videographers can save time by automatically blurring objects in a video, while also being confident in the accuracy of the underlying machine-learning algorithm.
The other technology, called Fully Homomorphic Encryption (FHE) Transpiler, enables developers to perform computations on encrypted data without being able to access personally identifiable information. Google believes that this tool can be useful for industries such as financial services, healthcare, and government, where it is critical to have a strong security guarantee for the processing of sensitive data.
Google has noted that privacy-enhancing technologies (PETs) are starting to become more widely used, after previously being mostly used in academic settings. The White House has recently expressed support for PETs, stating that they “will allow researchers, physicians, and others permitted access to gain insights from sensitive data without ever having access to the data itself.” Both the US and UK governments have held contests this year to develop PET solutions for addressing financial crime and public health emergencies.
