Trust What You See: How Truepic Authenticates Images and Videos In the Age of Deepfakes

This blog post was originally published at Qualcomm’s website. It is reprinted here with the permission of Qualcomm.

Truepic is leading the charge to make sure people know the origin of what they’re seeing online

Key Takeaways:

  • The ability to trust what you see when it comes to digital content is more critical than ever.

  • A host of different companies, from Google to Microsoft, have bought into the idea of authenticating photos and video, with Truepic helping to lead the charge.

  • Qualcomm Technologies has worked to optimize Truepic with the Snapdragon mobile platform, giving more Android phones these protections.

When images of Pope Francis strolling outside of the Vatican in a chic designer puffer jacket zipped around the internet in March 2023, it served as a stark reminder of how far fake content created by artificial intelligence (AI) had come. Unless you heavily scrutinized the picture of the pontiff — complete with blinged-out crucifix — you’d be convinced it was real.

The images went viral just as generative AI exploded in popularity. Since then, the tools to create realistic videos and images with simple prompts have gotten more sophisticated and accessible. It’s also meant that it’s harder to believe anything you see on the internet.

That’s where Truepic comes in. The company has led the charge to create a system that verifies through metadata whether images and videos are authentic, and is working with a broad set of partners, from tech titans like Microsoft and Google, to social media leaders such as Meta and software content providers such as Adobe.

This has allowed content creators, publishers and social media platforms to get more of handle on what’s real and what’s fake — which is critical as misinformation and disinformation online continues to be a problem.

In an interview with OnQ, Truepic CEO Jeffrey McGregor talks about the work his company is doing to restore trust in digital content.

Meet Truepic

Truepic came together in 2016 after a realization that there would quickly be a need for authentication and transparency in a world where fake content could harm industries ranging from commerce to online dating, and undermine elections around the world.

Truepic tested many approaches to authenticating content at the point of origin and spent years adjusting and evolving its product. Around the same time, different initiatives championed by the likes of Microsoft, the British Broadcasting Corporation (BBC), the New York Times, Adobe and Twitter began to emerge.

Truepic worked together with many of these companies to create and define a standard called C2PA, as well as a non-profit organization called the Coalition for Content Provenance and Authenticity (C2PA.org) to act as its steward. Amazon, Adobe, BBC, Google, Intel, Microsoft, Meta, OpenAI, Publicis Groupe, Sony and Truepic are among the steering committee members who now lead the group.

The C2PA standard relies on cryptographic hashing and digital signatures which are imprinted onto a photo or video the moment it’s created. The information includes data such as when the media was taken, the device or software that captured it, and whether editing software was employed. Another device running on C2PA can then unlock and look at the data to verify the image’s legitimacy.

That’s why it’s so important that Truepic works with so many players who support the standard.

While C2PA is an open specification, Truepic offers an enterprise-grade version of the standard with additional reliability, security and support. It also offers a separate software as a service called Truepic Vision that lets organizations request, capture, and review trustworthy digital photos and videos to replace slow and costly onsite inspections.

Fast forward a few years, and the level of sophistication and quality coming from AI-generated content has exacerbated the need for a company like Truepic.

“AI is not the problem,” McGregor said. “But it has certainly accelerated the need for a solution.”

Integrating with Snapdragon for on-device authentication

Qualcomm Technologies contacted Truepic to collaborate on a proof of concept that would eventually result in an authentication solution for the Snapdragon 8 Gen 3 mobile platform. Engineers on both sides worked daily to optimize Truepic’s services directly for use with the Qualcomm Technologies’ Trusted Execution Environment in the Snapdragon 8 Gen 3 mobile platform, providing vault-like security for a device’s sensitive information.

Having the Qualcomm Technologies’ Trusted Execution Environment handle the C2PA signing makes the authentication process even more secure because it’s even harder for bad actors to tamper with the cryptographic keys protecting the data.

McGregor praised Qualcomm Technologies’ support, noting that Truepic got attention and resources from the senior level to do a lot of the custom work required for the integration. While many of Truepic’s services run through apps powered by the cloud, both agreed it would be better if the authentication happened on the device itself. Having it built in just makes it easier for the end user.

“There’s a shared belief that easing adoption for smartphone OEMs by making this implementation turnkey and secure is a really, really important thing to do,” McGregor said. “Those shared values are what brought the collaboration together.”

What’s next

The relevance of Truepic’s services is best illustrated by the U.S. presidential election, which has already seen its fair share of AI-content muddle the race. It was one of 20 companies to sign the AI Elections Accord in February 2024, working alongside TikTok, Meta, Google, OpenAI and others to combat deceptive AI election content.

Companies, meanwhile, are increasingly looking at C2PA and Truepic because they don’t necessarily have the resources or time to deal with verifying images and other online content themselves.

“When you’re dealing with data integrity, trust and transparency, the details and the nuances matter, especially at this scale,” McGregor said. “You want to get this right, and relying on a team that has spent significant time working in the field and understanding the technical and security nuances has been very valued by our customers.”

More companies are also keen to sign on because they see where regulations around AI safety are headed. Many see C2PA as a way to get in line with rules that are coming down the pike.

Truepic itself is focused on getting its technology embedded into the next generation of flagship phones running on the Snapdragon 8 Gen 3 platform. McGregor said that will allow Truepic to achieve a new level of scale for the technology.

“It’s the natural next phase of the journey,” he said.

Learnings from Truepic

  • AI is not the problem, but it has certainly accelerated the need for a solution.
  • Government regulations around AI are emerging, pushing companies to look at standards like C2PA.
  • Businesses don’t want to divert attention away from running their core business, which makes Truepic’s turnkey Enterprise C2PA solutions attractive.

Here you’ll find a wealth of practical technical insights and expert advice to help you bring AI and visual intelligence into your products without flying blind.

Contact

Address

Berkeley Design Technology, Inc.
PO Box #4446
Walnut Creek, CA 94596

Phone
Phone: +1 (925) 954-1411
Scroll to Top