It’s Deepfakes vs. Deepfakes: Google releases a Huge Database

Google recently released a huge database that includes deepfakes coming from 3,000 AI-generated videos made by the publicly available algorithms.

Generative algorithms have been so good in producing deepfakes by media synthesis that you hardly distinguish between real and fake. Such fake images and videos are not easily identified through computer software as well, therefore, experts are now working towards introducing a better option to cater to this.

Just recently, Google came out with 3000 AI-generated videos that could be used for manipulation. Google is working on its deepfake detection tools, which is the prime reason why they released an open-source database of deepfakes.


Google collected these manipulative videos after working 28 actors who recorded videos while performing certain tasks and then deepfake algorithms were run on them.

Facebook has also reported that they would be releasing a certain database as well.

Using AI to identify deepfake images can be used to train and test the automated detection tools. Identifying these deepfakes before US elections will also be great to eliminate the chances of rigging. However, technical measures are not enough to identify deepfakes, certain actions on political, legal and social grounds should be taken as well to accomplish this.



Read next: Now companies are getting scammed through AI deepfakes and we are worried

No comments:

Post a Comment