End-to-end standard pushed to out deepfakes and other misinformation

A proposed open standard for detecting deepfakes has been published by a broad coalition of companies involved with artificial intelligence, content creation or authentication.

The Coalition for Content Provenance and Authenticity is pushing a standard for software and hardware used to create, edit, manage and verify content.

Creators would be able to certify what they did in generating a still image or video. They also would be protected when alterations are made downstream by certifying those unauthorized changes.

Editors, distributors (such as social media platforms) and consumers of digital content could largely allay worries about AI-generated misinformation, according to the coalition’s standard specifications.

But law enforcement, court officials, regulators, celebrities and even human rights workers could also benefit from the ability to investigate the provenance of allegedly true content.

Some coalition members, such as the United Kingdom’s BBC, CBC/Radio-Canada, The New York Times and Twitter, have everything to lose if they are distrusted. Others, like vendors Adobe, Niko, Serum of Truth and Truepic, hope to benefit from societies seeking more information certainty.

(Truepic has been working since at least 2020 with U.S. wireless vendor Qualcomm Technologies to make Truepic provenance software native on phones using Qualcomm chipsets.)

Many entities faced with a reality dysfunction created by deepfakes want to find a reliable farm-fertilizer detector and will invest in any technology solution with a shot at working.

This coalition is not even the first to take a crack at solving this problem. In 2020, a group similar to this one was trying to get its hands around deepfakes.

Ultimately, though, the potential authentication standard will not be the key piece of the puzzle. Information distributors still have to have the integrity to test content. And information consumers have to be honest and curious enough challenge content that bolsters their own uninformed biases.

Article Topics

AI  |  authentication  |  biometrics  |  deepfakes  |  fraud prevention  |  spoof detection  |  standards