Battling the Deep Fake Surge:Reality Defender Keeps it Real

The startup leading the fight against AI-generated disinformation is exploring opportunities with NBCUniversal following the Comcast NBCUniversal LIFT Labs Generative AI Accelerator.

A fake image of a fire at the Pentagon swept across Twitter. News organizations reported it as true. The stock market dipped. Panic set in.

Though persuasive to the layperson, the image was a fake — created by artificial intelligence. The accessibility of AI image creation tools via web and mobile applications paves the way for the production of convincing “deep fakes” making it easy for bad actors to create images, videos, and text messages that appear real.

On day one of the recent Comcast NBCUniversal LIFT Labs Accelerator, Ben Colman, the Co-founder and CEO of Reality Defender, was invited on CNBC’s Last Call with Brian Sullivan to shed light on the falsehood. Reality Defender is a deep fake detection platform with sophisticated technology to verify the authenticity of audio, video, images, and text.

“Today, anybody can access a fake face or voice generator — creating potential for something incredibly entertaining or something incredibly dangerous. The emergence of generative AI gives non-technical people the ability to easily create a media ‘virus’ that can spread quickly across the internet.”

— Ben Colman

Think of Reality Defender like the antivirus software aiding companies in the detection of deep fakes. In its early days, the company provided its technology to more than 100 government agencies around the world for free, helping them scour the web and dark web for misinformation and disinformation. Today they work with heavyweights like Visa, Microsoft, the North Atlantic Treaty Organization (NATO), and others. It also helps banks with real-time detection of counterfeit voices or texts used to enable fraudulent transactions.

The company is also leading the charge for government regulations that safeguard individuals and organizations against deep fakes. Colman helped launch Deep Trust Alliance, a 501(c)(6) nonprofit lobbying U.S. Congress to enact laws for deep fake protection, particularly advocating for social media and online platforms to identify, block, and report AI-generated content.

The implications are clear. People need to know what is real and fake.

“We inherently trust what we see and hear. The challenge is that everything we see online and everything we hear on the phone can be faked,” he said. “We can flag potential fakes in real time — in less than a second — allowing organizations to prevent potential fraudulent content from going viral.”

Joining Forces with LIFT Labs

Reality Defender recently completed the Comcast NBCUniversal LIFT Labs Generative AI Accelerator. The six-week program fostered connections with executives across Comcast NBCUniversal and Sky, along with programming on topics such as enterprise sales, the future of generative AI technology, and media training.

“We were blown away by the breadth of experience offered by the LIFT Labs Accelerator,” said Colman. “With LIFT Labs, we have access to a wide array of global companies within the larger Comcast NBCUniversal Family. Plus, LIFT Labs will be our internal champion for the long haul, not just during the Accelerator program.”

As a result of the Accelerator, Reality Defender is piloting their technology with NBCUniversal to explore solutions for securing their outbound content in the public space.

Joining the Accelerator propelled Reality Defender’s mission to equip authorities and enterprises with the tools to keep pace with the fast-evolving deep fake technology. Getting it right is critical to the future of human interaction.

“We are going to be on a video call and not realize the person, the face, the voice is completely fake because it communicates with us perfectly in real time,” said Colman. “We’re trying to get ahead of this problem right now and build technology to detect it and also establish the requirements for platform providers to address what could become the largest problem for our generation online.”

A Forward Vision

Colman is particularly invested in this cause, eager to ensure the safety of his children who are already interacting with the online world.

“To them, the online world is as tangible as the printed newspapers were for me. We aim to educate them about potential fraud and the importance of verification so they can understand what’s real and fake,” he said.

Despite his company’s mission, Colman is a generative AI enthusiast who appreciates its transformative power, as long as people can tell what’s been generated by AI.

“At Reality Defender, we absolutely love and embrace generative AI,” said Colman. “We think that it’s useful, it’s helpful, and it’s the future. We just think that viewers should be aware that something was generated by AI. Transparency in discerning what’s real and what’s fake is critical for everyone.”