Google recently changed its platform policy to prohibit developers from using the services of its Colaboratory platform for research on the Deepfake project, ending the large-scale use of the platform's resources for this purpose.
Colaboratory, or Colab for short, is a product developed by the Google Research team. In Colab, anyone can write and execute arbitrary Python code through a browser. It is particularly suitable for machine learning, data analysis, and educational purposes. Technically, Colab is a hosted Jupyter Notebook service. Users can use it directly without setup and also get free access to computing resources such as GPUs to accelerate project research.
Colab supports most major browsers and has been fully tested on the latest versions of Chrome, Firefox and Safari.
Deepfake can be trained to swap faces on video clips, adding realistic facial expressions to make the final video effect look realistic for the purpose of faking it. The technology itself is based on machine learning and is currently a popular area of research in the industry, but it has been used by a number of people to spread fake news and fake political or pornographic videos. As this technology is ethically controversial, many people and businesses are very concerned about it.
According to archive.org's web page historical data, the Google earlier this month modified Colab's banned items list, adding Deepfake to the blacklist (red box below).
If a developer attempts to continue training a Deepfake project on the Colab platform despite the blacklist, the developer will receive an error like the following.
You may be executing code that is not allowed, which may limit your ability to use Colab in the future. Please be aware of the prohibitions we have set out in the FAQ.
The impact of this new restriction is expected to have a profound effect in the Deepfake space, as many users take advantage of Colab's pre-trained models to start their projects. And Colab itself is so easy to use that even those with no programming background can get started quickly, which is why a very large number of Deepfake tutorials recommend using Colab.
It is not known whether Google is enforcing this policy for ethical reasons or as a limitation on the excessive load of computing resources utilized by these projects.