Skip to main content

This crazily realistic video forgery of Obama was generated by a lip-syncing AI

Teaser -- Synthesizing Obama: Learning Lip Sync from Audio
Regardless of which side of the political aisle you sit on, chances are you’ve got some strong opinions on “fake news.” Whether it’s comments taken out of context, or quotes being outright fabricated, fake news is a frustrating byproduct of today’s twenty-first century news cycle.

Well, we’re sorry to tell you that things are about to get much, much worse!

At least, that’s based on a frankly crazy demonstration of artificial intelligence carried out by computer scientists at the University of Washington. Using a cutting-edge artificial neural network, they’ve developed an AI that’s able to produce new video footage of former President Barack Obama speaking, which perfectly matches recorded audio of him.

“We developed an algorithm that can generate a believable video of Obama from his voice, based on a recurrent neural network that learns how to do this by analyzing hours of Obama’s weekly address footage,” Dr. Supasorn Suwajanakorn, a researcher on the project, told Digital Trends. “Unlike prior work, we never require the subject to be scanned or a speech database that consists of videos of many people saying predetermined sentences. We learn this from just existing footage. This has the potential to scale to anyone with minimal effort.”

Image used with permission by copyright holder

So with that being the case, why did the researchers choose the likes of Barack Obama to carry out the jaw-dropping tech demo? No, it’s not for partisan political reasons. “The technique we used — deep learning — requires lots of data,” Suwajanakorn continued. “And this dataset is well suited because it’s large: over 20 hours, easy to collect, contains only Obama in high-res, and public-domain, which is free for researchers to use.”

It’s worth noting that the researchers aren’t creating a CGI’d version of Obama from scratch. Instead, they’re doing a more sophisticated version of what the filmmakers behind a movie like Forrest Gump did with archival footage: taking existing video and then editing it to fit new dialog. In this case, that means moving Obama’s mouth to match what he’s saying in the audio — all while incorporating everything the AI has learned about his unique mouth movements to make it appear authentic.

We’re intrigued — albeit disconcerted — to see what happens when someone mixes the University of Washington’s research with this previous project, using a neural network to mimic the voice of (among others) President Obama.

See what were we saying about the future of fake news?

Editors' Recommendations

Luke Dormehl
I'm a UK-based tech writer covering Cool Tech at Digital Trends. I've also written for Fast Company, Wired, the Guardian…
Get ready to waste your day with this creepily accurate text-generating A.I.
ai spots writing by fake news feature

Whether you believe it was one of the most dangerous versions of artificial intelligence created or dismiss it as a massive unnecessary PR exercise, there’s no doubt that the GPT-2 algorithm created by research lab OpenA.I. caused a lot of buzz when it was announced earlier this year.

Revealed in February, OpenA.I. said it developed an algorithm too dangerous to release to the general public. Although only a text generator, GPT-2 supposedly generated text so crazily humanlike that it could convince people that they were reading a real text written by an actual person. To use it, all a user had to do would be to feed in the start of the document, and then let the A.I. take over to complete it. Give it the opening of a newspaper story, and it would even manufacture fictitious “quotes.” Predictably, news media went into overdrive describing this as the terrifying new face of fake news. And for potentially good reason.

Read more
A.I. is getting scary good at generating fake humans. Watch this demo
japanese gan people dont exist screen shot 2019 05 03 at 16 23 06

【データグリッド】全身モデル自動生成AI | [DataGrid] Model generation AI

A.I. is getting pretty scarily good at lying to us. No, we’re not talking about wilfully misleading people for nefarious means, but rather creating sounds and images that appear real, but don't exist in the real world.

Read more
A.I. can generate pictures of a finished meal based on just a written recipe
tel aviv ai generates food pictures recipe book preparing

Recipe books contain photos. You know why? Because most of us aren’t capable of reading through a list of ingredients and associated instructions and picturing exactly what is supposed to come out of the oven or other cooking appliance at the end of the process. Artificial intelligence, it seems, doesn’t have quite the same problem -- at least, not according to a project carried out by researchers from Tel-Aviv University in Israel.

Using a training data set of approximately 52,000 written recipes, along with images showing the completed foods, the researchers were able to devise a system that can read a recipe and then generate a picture showing what the end result is likely to look like.

Read more