Meanwhile, as deepfake technology progresses, concerns about its potential misuse also heightens. (Photo: AFP)

Until now, AI-generated deepfakes were fooling people to believe the existence of certain individuals. Now, researchers have generated realistic portraits which can also sing!

As The Verge reports, researchers from Imperial College in London and Samsung's AI research center combined a single photo and audio file to generate a singing or talking video portrait using machine learning.

For instance, a real clip of Albert Einstein speaking combined with his photo can be turned into a never-seen-before lecture, or Grigori Yefimovich Rasputin singing Beyonce classic 'Halo', all using a single picture and an audio file. Meanwhile, as deepfake technology progresses, concerns about its potential misuse also heightens.