Google banned training of deepfakes in Colab

Google banned training of deepfakes in Colab

San Francisco, 6/1/2022

Google has banned training of deepfakes in Colab by changing its terms of service for its users. The company has not publicly announced the change.

Google made the changes in the last two weeks, ending the use of its platform for this purpose. Colab is an online computing resource that allows researchers to run Python code directly from the browser, using free computing resources, including GPUs, for their projects.

The platform was used to train deepfake AI models that created fake videos, articles, images and other types of content that fooled people into thinking they were real.

The full list of discontinued features includes:

File hosting,

media serving or other web service offerings that are not related to interactive

processing data with Colab

downloading torrents or participating in peer-to-peer file sharing

the use of a remote desktop or SSH

connecting to remote proxies

mining cryptocurrencies and carrying out denial-of-service attacks

cracking passwords using multiple accounts to bypass access or resource usage restrictions, and

Creating deepfakes.

It is not known if Google made the decision based on a specific case of misconduct or if the company recognized the potential for abuse of this project.

Deepfakes are synthetic media in which a person in an existing image or video is replaced with the likeness of another person. The most recent example was a deepfake by Elon Musk, who promoted a crypto project and managed to scam people out of thousands of dollars.