Fake news is a major and prickly subject nowadays and there is no sign that it’s leaving at any point in the near future. Truth be told, the situation may deteriorate before it shows signs of improvement. Innumerable individuals are as of now effortlessly hoodwinked by genuine looking content and professionally photoshopped pictures. Be that as it may, imagine a scenario where they can be tricked by video too. While the machine learning algorithms created by University of Washington specialists have more grand objectives as a main priority, this “lip-syncing tech” could likewise be utilised to trick and deceive individuals.
Envision taking clips from one video, say a meeting of previous US President Barack Obama, and blending it with sound from a totally unique video, similar to an entirely unrelated interview. While actually conceivable, the outcomes are not usually all that realistic. Utilising this new machine learning method, University of Washington scientists possess the capacity to deliver high-quality audio and video syncing, to the point that shockingly resembles the genuine article.