Episode 37 – From Detection to Provenance

As generative AI makes it increasingly difficult to know whether a voice, image or video is authentic, the question of digital trust is shifting. Detecting what is fake may no longer be enough. We may need a reliable way to prove what is real, where it came from and who — or what — created it.

In this episode of Innovating Trust, Steve Atkins speaks with Slavko Kovačević, Co-Founder and CEO of DeepMark, about digital provenance, AI watermarking and the challenge of building trust into synthetic media.

They discuss why deepfake detection can become a technological arms race, how invisible machine-readable identifiers can travel with digital content, and why provenance could become an important part of the infrastructure supporting AI-generated media. The conversation also looks at the growing regulatory pressure around synthetic content, the practical challenges of making watermarking survive real-world editing and distribution, and what happens as AI-generated voices move into areas such as banking, customer service and communications.

As the boundary between human and synthetic content becomes harder to see or hear, trust may increasingly depend not on recognising a fake, but on being able to verify the source.

Listen to ‘Innovating Trust’ on your regular Podcast platform;

podcasts.apple.com/gb/podcast/the-quantum-spac…

open.spotify.com/show/4S5GpoVMWzCVGz8bk3xJYG?u…

music.amazon.de/podcasts/09dccee7-168c-4a15-ab…

music.youtube.com/playlist?list=PL5BBxCBd2lp_R…

pocketcasts.com/podcast/the-quantum-space-inno…