Re:NeRF: An eventual successor for deepfakes?
by wongarsu on Monday June 27, 2022 @12:18PM
(#31896738)
80% of the problem could be solved with a reliable signature scheme that allows some remixing of video content. So if CNN publishes a video, signed with their key so it's verifiably CNN, we need the ability to take a 20 second bit of it and still have a valid key attached that verifies that the source is CNN (without trusting the editor). Then you can share clips, remix it, etc, and have integration in social media that attests the source.
Re:NeRF: An eventual successor for deepfakes?
by intrasight on Monday June 27, 2022 @01:13PM
(#31897571)
My plan to solve this "20 second bit of it" is that it's done at the analog hole. Whatever is painting those pixels, a smart TV for instance, will be coordinating with cloud services to fingerprint at a relatively high temporal resolution - maybe 5 seconds. The video itself is the signature. But we will need either trusted analog hole vendors or some trusted non-profit organization - or likely both. I think that "viewing" will be delayed by perhaps 30 seconds to allow for that signature analysis. These smart TVs will overlay a scorecard for all displayed content, and owners will be able to set device scorecard thresholds such that low-scoring content will be fuzzed out.
I sincerely hope this dystopian vision of the future is satire, but it's already a worrying sign of the times that I'm not sure.
Technically, we could try and build a trusted computing-like system to ensure trust from the sensor all the way to a signed video file output, but keeping that perfectly secure is likely to be virtually impossible except in narrow situations, such as CCTV installations. I believe Apple may be attempting to do things like this with how Face ID is implemented on iPhone, but I suspect we'll always find ways to trick any such device.