I agree that the technology is great, and it will empower small creators, but I'm also worried about the cowboy behaviour of all these tech billionaires.
If you record my voice at a conference and then create a synthetic replica.. why would I care? You didn't make me do any additional work or anything
In the case of a pornographic video there is no issue if it's clear from the context or content that it isn't actually me doing what's in the video
When you talk to ChatGPT, I don't actually think Scarlett Johansson is speaking to me
If I make a fake phone call recording with her synethetic voice and claim it's real and it somehow hurts her then that's an issue - but that's a different legal matter entirely
My reaction would never be "we must make it so people can't do deepfakes anymore". That would cause people to stop using it for positive/benign things as well. If someone is spreading deepfake porn of someone, and you could make the case that they are doing so in order to harm that person's reputation, then legal action would make sense, I think.
I'm not buying that people here are "fine" with this. This is one of those things people might be fine with, until they find themselves in that exact spot.
Now, what kind of people will find themselves in that spot? Celebrities, obviously.
One thing is being used in material that will defame your character (spam, fraud, porn, whatever) - another thing is to be used in material that will potentially take away your livelihood.
If someone clone, say, Tom Cruise - and makes a movie with his digital twin, he sure as shit is entitled to royalties for that. People go to see the movie because they think it is Tom Cruise, not because it's some generic AI avatar of him.
That's a deepfake problem but not porn problem. "Deepfake == porn" characterization is not helpful if those are the real problems.
I'm gonna suggest that people who are blase about this issue are comfortable in the knowledge that it will never affect them. HN contributors might have 99+ problems, but being lusted over by the internet at large isn't one of them.
So you basically are saying that in these cases law system should have a precedent against deepfakes so that you would be to able to argue on some basis against deepfakes made off of you.
The people now getting concerned about this are setting those precedents, so that when shit hits the fan in your life (it probably won't), you will have an easier path in the court.
(No no, you're perfectly right [except perhaps about "the technology is great, and it will empower small creators"], but yagotta admit, your example in justaposition with your user id is funny.)
Are there actually any "benign" uses for deepfakes?