Visual 1st Spotlights: Re-establishing Trust in Visuals

Untangling the potential solutions – from image content credentials to digital certificates, forensic watermarks, deepfake detection, and beyond.

The pervasiveness of AI has radically altered the dynamics of trust in imaging. We’ve gone from time-and-skill consuming cut/paste/rephotographing in the silver halide days, to generating photo-similar representations of anything we can dream up in seconds and on demand.

There’s broad industry – and even political – consensus that the issue of veracity in imaging is critical for both commercial and societal reasons, but we’re still early on in developing solutions, and there’s much confusion on the potential and limitations of the various approaches ranging from metadata credentials, to watermarking, to deepfake detection.

In this session we’ll examine what’s being done:

• To assist content creators to certify that their images are authentic and can be trusted, including with respect to IP.
• To enable imagery viewers and buyers to ascertain whether the images they’re viewing, or are about to publish under their responsibility, meet the necessary authenticity requirements.

• Paul Melcher
• Jeffrey McGregor
• Eric Wongrowski

Moderators:

• Hans Hartman
• Alexis Gerard

Visit Event Website