PCMag editors select and review products independently. If you buy through affiliate links, we may earn commissions, which help support our testing.

Latest Deepfake Tech Will Have You Dancing Like Bruno Mars

Everybody dance now? Researchers at the University of California, Berkeley explore how to transfer the motion of one person in a video to another.

 & Michael Kan Principal Reporter

Our team tests, rates, and reviews more than 1,500 products each year to help you make better buying decisions and get more from technology.

Our Expert
LOOK INSIDE PC LABS HOW WE TEST
65 EXPERTS
43 YEARS
41,500+ REVIEWS

Can't dance? There's AI technology for that.

Researchers at the University of California, Berkeley have found another way artificial intelligence can be used to create fake videos: making you look like a professional dancer.

The system can train itself to render lifelike images of anyone performing the dance moves of pop star Bruno Mars or a ballet dancer.

The system works by essentially capturing the poses and motions of the professional dancer and then mapping them on to the body of the amateur, the researchers wrote in their paper, which was released this week. The system first does this by breaking down images of both the professional and amateur into stick figure-like images. This is done so it can easily read the movements of one body and then puppet the other to follow along.

The technology will then create some image fakery with the help of an AI-powered algorithm called generative adversarial networks (GANs), which have proven to be adept at creating photorealistic images of people.

UC Berkeley Dance

GANs essentially work by pitting two neural networks against each other to complete a certain goal. This can be effective when it comes to image generation; one network will be assigned to synthesizing the images, the other will act as a "discriminator" and discard images that look flawed.

"These two networks are trained simultaneously and drive each other to improve, as the generator must learn to synthesize more realistic images to deceive the discriminator," the UC Berkeley researchers wrote in their paper.

In their experiments, the researchers filmed about 20 minutes of real-time footage of their amateur dancers performing a range of movements. The AI system then had enough data to begin synthesizing fake video sequences. One GAN was also incorporated to specifically render realistic faces and expressions to go along with body movements.

UC Berkeley Dance 2

The results aren't perfect. For instance, the system struggles to replicate the movements of a ballet dancer performing kicks and spin moves. Nevertheless, the technology can certainly be used for amusing effect.

The researchers titled their paper "Everybody Dance Now," alluding to the almost limitless possibilities of what AI-powered video generation can achieve — for better or worse. Others have used the technology to create porn that's been modified with the faces of celebrities without their consent, often referred to as deepfakes.

The UC Berkeley researchers noted their own system was developed without the need for "expensive 3D or motion capture data." Time will tell if their demonstration will lead to an explosion of fake dance videos.

About Our Expert

Michael Kan

Michael Kan

Principal Reporter

My Experience

I've been a journalist for over 15 years. I got my start as a schools and cities reporter in Kansas City and joined PCMag in 2017, where I cover satellite internet services, cybersecurity, PC hardware, and more. I'm currently based in San Francisco, but previously spent over five years in China, covering the country's technology sector.

Since 2020, I've covered the launch and explosive growth of SpaceX's Starlink satellite internet service, writing 600+ stories on availability and feature launches, but also the regulatory battles over the expansion of satellite constellations, fights with rival providers like AST SpaceMobile and Amazon, and the effort to expand into satellite-based mobile service. I've combed through FCC filings for the latest news and driven to remote corners of California to test Starlink's cellular service.

I also cover cyber threats, from ransomware gangs to the emergence of AI-based malware. In 2024 and 2025, the FTC forced Avast to pay consumers $16.5 million for secretly harvesting and selling their personal information to third-party clients, as revealed in my joint investigation with Motherboard.

I also cover the PC graphics card market. Pandemic-era shortages led me to camp out in front of a Best Buy to get an RTX 3000. I'm now following how the AI-driven memory shortage is impacting the entire consumer electronics market. I'm always eager to learn more, so please jump in the comments with feedback and send me tips.

The Best Tech I've Had:

  • My first video game console: a Nintendo Famicom
  • I loved my Sega Saturn despite PlayStation's popularity.
  • The iPod Video I received as a gift in college
  • Xbox 360 FTW
  • The Galaxy Nexus was the first smartphone I was proud to own.
  • The PC desktop I built in 2013, which still works to this day.

Read full bio