Google appears to have banned deepfake projects and their training from its Colaboratory (Colab) service , putting an end to the misuse of the platform for this purpose. 's resources
Google's Colab service is an online computing resource that allows researchers to run Python directly through the browser, while using free computing resources (including GPUs) to power their projects.
See also: Work from home: Google contractors don't have the same privileges as employees
Due to the multi-core nature of GPUs, the Colab service is ideal for training machine learning projects such as deepfake models or performing data analysis.
Deepfakes can be trained to change faces and voices in a video . The technology helps add realistic facial expressions and movements to make the result look more genuine. This means that someone could exploit this technology to create a fake video that shows a person talking, when in reality it is not.
Such videos have often been used to promote crypto scams, spread fake news, create revenge porn, or simply for entertainment. However, the lack of ethical restrictions on their use has caused great concern.
See also: Google Drive: Gets copy/paste shortcuts
Google: Bans deepfake training in Colab service
Based on archive.org data, Google banned deepfake projects in Colab earlier this month, placing them on the list of disallowed projects.
As noted on Discord by DFL developer “chervonij”, those attempting to train deepfakes on the Colab platform currently see the following message:
“You may be running code that is not allowed and this may limit your ability to use Colab in the future. See the prohibited actions specified in our FAQ“.
This move by Google will certainly have a significant impact on the deepfake world, as many users use pre-trained models with Colab to jumpstart their high-resolution projects . Until now, Google's Colab service has made this process very easy even for those without coding knowledge.
Google: Stop resource abuse
It is not known whether Google banned deepfakes in Colab due to ethical concerns or due to the continued abuse of free computing resources for these projects.
According to the company, Colab should be used by researchers who need energy that costs many thousands of dollars to achieve scientific goals.
See also: How to use comparison operators in Google Sheets
However, there have been numerous reports that some users are exploiting the platform's free tier to create deepfake models on a large scale. This has taken up a significant amount of Colab's available resources for extended periods.
The list of disallowed projects includes the following:
- file hosting, multimedia viewing, or other web service offerings not related to interactive computing with Colab
- using remote desktop or SSH
- downloading torrents or participating in peer-to-peer file-sharing
- connecting to remote proxies
- cryptocurrency mining
- password cracking
- executing denial-of-service attacks
- using multiple accounts to overcome access or resource usage restrictions
- deepfakes creation
Source: www.bleepingcomputer.com
