Can A Metaverse AI Win America’s Got Talent? (And What That Means For The Industry)

When “Simon Cowell” performed on the America’s Got Talent stage, no one was more shocked than the man himself. How did this happen? In their one-of-a-kind audition, leading AI company Metaphysic transformed singer Daniel Emmet into Cowell and earned a spot in the semifinals.

Founded by visual FX and AI expert Chris Umé and serial entrepreneur Tom Graham, Metaphysic is pioneering the future of hyperreal, AI-generated content through mainstream entertainment. Viral internet sensation DeepTomCruise put Metaphysic on the world stage for its incredibly lifelike videos featuring Cruise and other celebrities like Paris Hilton and Emma Chamberlain.

To bring “Simon Cowell” to life for AGT, Chris Umé used the same precision animation and proprietary technology for a unique live performance that had not been done at this level.

As we approach the semifinals, I had the opportunity to speak with Tom and Chris from Metaphysic and learn how they plan to use this technology to transform the entertainment industry.

What Is Synthetic Media?

Synthetic media is the rapidly-expanding world of digital experiences and objects generated with input from artificial intelligence.

At Metaphysic, they are pioneering hyperreal synthetic videos that are highly realistic and often indistinguishable from digital content that is captured on cameras.

For the Tom Cruise deepfake videos, for example, Chris spent nearly three months training an AI model on images and videos of the real Cruise, capturing him from multiple angles and in a variety of lighting conditions.

Then Chris shot base videos using a body double, then generated the deepfake videos by combining the body double footage with actual video of Cruise’s face.

Chris and Tom followed a similar process for their performance at America's Got Talent, and their efforts earned them a spot in the semifinal round. The team plans on continuing to wow the AGT audience and judges with another groundbreaking performance.

What the Future of Synthetic Media Holds

When I asked Chris and Tom for their thoughts on what the future of synthetic media might look like, Tom said:

“We are going to participate in [the internet] in a much more profound way, and it's not going to look so much like the cartoony metaverse that we see today. We will be doing everyday things, like video conference calls, inside a more immersive virtual world. We could create hyperreal synthetic versions of regular people like us, including our voices and faces, and we are going to interact in a way that we can consume and watch and interact…to generate a human experience in the metaverse that is much more substantial and rewarding that what we get today.”

Tom believes that in the future, as much as 80% of the time we spend online will be in these types of synthetic content interactions that will feature regular people.

Chris also shared that he believes we will be able to enjoy real connections in the metaverse using this technology. He imagines a time when he might be able to have a virtual “live” breakfast with his grandparents, who live on the opposite side of the world, and be able to enjoy a true experience of being in the room with them and talking to them. Then 20 years from now, he could revisit that memory or even share it with his children, so they understand what his grandparents are like and what it’s like to interact with them.

When asked to predict the future of the entertainment industry with synthetic media, Chris said:

“I think we’ll have two lives: One will be here on earth, how we walk and talk and eat, and the second life will be the one in the metaverse…I hope to bring people together in ways, which we couldn't do before — even if it's not physical, even if it's by bridging language barriers. Because with the content I create and which we are creating with the company as well, I also hope to keep entertaining people and to keep being creative with the tools that we have and to try out new things that people don't expect.”

Metaphysic has built a tool called Every Anyone that people can use to create their own hyperreal likenesses and safeguard their private biometric face and voice data. People who create hyperreal avatars can use them anywhere, including social media, online courses with classmates, video games, and more. This tool is part of Metaphysic’s plan to scale synthetic media and make it more accessible to regular users.

Ethical Considerations of Hyperreal Synthetic Media

As with all innovative technology, there are ethical concerns.

“What we hope doesn’t happen,” Tom says, “is that large corporations or bad actors use [your] data to create versions of you without your consent, your permission, or without your direct interaction. What we're trying to create with Every Anyone is a way for individuals to own and control your data, the data that you bring into the internet and use to train content creation. The algorithm should belong to you, and you should be able to delete it.”

Metaphysic is committed to doing the work to make sure the technology they bring to the world is safe, deliberate, and considerate.

When I asked about the ethics of hyperreal synthetic media, Tom said the company is doing what it can to educate the public. He shares:

“It's what we can do to raise awareness among regular people that media can be manipulated, and maybe you should take a beat before you send that [content] on or before you see some kind of video online and you become outraged or have a particular emotional reaction.”

What Comes Next

Part of the mission of Metaphysic is to work with clients and the general public on the ethical use of deepfakes. In the meantime, their team will continue to astound us with their incredibly realistic content.

Watch Metaphysic pioneers Tom and Chris continue their journey in the semifinal round of America's Got Talent!

Watch my whole interview with Tom and Chris here: