The danger is not in false positives, but false negatives. The very existence of this kind of things erode trust and sow paranoia.
A simple morph cut in a John Pilger interview of Assange made a sizeable portion of nutjobs believe Assange has been long dead. Don't think this kind of behaviour can't eventually extend to the mainstream.
Cryptographic signing of video footage is a useful blockchain application, but it will presumably be subject to the same flaws as domain security certificates.
If you can sign a video then you can sign a doctored video. If it's only your camera that can sign the video, not you directly, don't be fooled that it will be possible to protect the private keys in the camera from extraction.
>If you can sign a video then you can sign a doctored video.
I don't understand your point, you wouldn't sign a doctored video unless you wanted to do so. It's entirely possible to apply the principles of PGP to video.
>If it's only your camera that can sign the video, not you directly, don't be fooled that it will be possible to protect the private keys in the camera from extraction.
I doubt this would be the solution on which the world settles.
When you publish a video of you speaking at a public event, you sign it. When someone else publishes a doctored video of you, you do not sign it.
This alone doesn't protect you in the case that someone speaks and then later intentionally doctors and signs the video in order to change what they said. In this case trusted third parties (e.g. news organizations) could sign videos as well. A set of signatures taken together can provide trust.
Signing video would prove it existed in that form at a particular time. Derivative instances could be linked back to the original. Every modification is like a transaction. You would always be able to follow a piece of media back to its original published source. Just imagine any given media stream as an edit decision list containing clips, each of which points back to an asset, each asset being a piece of video with a date of publication.
The aset isn't a single entity in one place, it can be distributed via IPFS or whatever. The earliest known version of a thing is the canonical one for practical purposes. In this view blockchain isn't producing coins for hoarding, but tags for people to locate things on the public graph.
Recording a hash of a video in the Bitcoin blockchain creates public evidence that the video existed at a specific point in time. If that point is very soon after the events portrayed in the video then it can increase confidence in the video's authenticity.
I think it would be more like prior art, to distinguish originals from copies rather than to authenticate the content itself. Given two pieces of similar-but-different video the one with the earliest timestamp is presumably the original.
A simple morph cut in a John Pilger interview of Assange made a sizeable portion of nutjobs believe Assange has been long dead. Don't think this kind of behaviour can't eventually extend to the mainstream.