This is my friend Paul, first speaking in English. Then HeyGen, generates a video of him speaking in French.
record a video of at least 30 seconds
go to HeyGen.com
create a free account. you get 2 free credits, enough to translate 2 videos
login and click Labs in the lower left corner to get to the latest features
click Video Translate
drop in your video, choose a language, click Submit
My attempts kept hitting “Failed,” prolly because HeyGen is backlogged with nearly 30,000 videos in queue for translation at the moment. I reported it in the support chat. They completed the conversion for me and emailed me the link.
Evidently they have an upcoming webinar
That is some scary stuff, IMO.
Haven’t read the article yet (will do so later today), but my first concern is “How long before we get ‘doctored’ videos being used to sway public opinion and then being used for court testimonies?”
They will require actual attendance, I’m thinking, to provide testimony.
Doctored photographs, for instance, have been in existence for a very long time, and there are methods to get photos into the record.
Something will get developed, I suppose. But deepfakes have been around for a bit.
The courts don’t scare me so much as our politics.
I think courts will have issues but will muddle through.
I’m ready to see videos of Biden on a “hot mic” admitting to helping Hillary Clinton procure babies to drink their blood for adrenochrome. Or Fauci discussing how he worked with Obama to spread COVID as a smokescreen for why Biden seemed to win the election that Soros rigged.
“There’s plenty of evidence for all of that; we just need to create it.”
People will unironically feel that way. We all know it’s true, I just need to spread the word, it’s ethical if it’s true anyway.
Yes, people differ in their opinion of the ethics of deceiving others for the purpose of getting them to do what they consider the right thing. I don’t think anyone is unequivocally for or against it. For example, you’re not going to tell a small child the rationales for following all the rules. OTOH, people who would use deepfakes to sway an election probably wouldn’t be so sanguine if it benefited the opposing party.
They have, though they’ve sort of sucked and require time and money to make. Same with signatures, photographs, audio…
I think we’re heading for a future where anyone can deep fake of anything instantly and easily.
I’m not sure how we’ll deal with that at all.
Live events will become more popular, I suppose.
All recordings are now potentially fiction.
Everyone also gets to claim that the recording of them saying or doing something they shouldn’t be doing is fake. You can’t trust eyewitnesses and you can’t trust the crime being captured on video or film.
I anticipate it won’t take long to fake live-action events that can react to input (like tell us the date, do a push-up, etc.)
How long until it can be made to fool the viewer without massive development on a single example? That’s the question… 5 years? 15? Of course it’ll get more sophisticated as it goes, but I wouldn’t be mindblown if in 5 years, you could have a digitized person on a monitor seeming to be IRL and reacting according to prompts given.