I agree that it seems hard to find a positive use for this kind of thing. What I could envision happening is essentially the end of the video testimonial for product reviews. Nobody will ever believe a "busy mom's" TikTok style review for an egg-whisker, skin cream, hair supplement etc again.
It will eventually make consumers extremely cynical of everything they're ever shown and basically break advertising.
Sadly, I don't think we can count on the average consumer showing that level of discernment.
Case in point: All the Facebook pages sharing clearly AI-generated images of "dream houses" and "my 3-year-old built a giant sand sculpture of Jesus all by himself," and the thousands of likes and comments they get.
Not to mention all the traditional hoaxes, scams, etc. that continue to exist and thrive even in the Internet age.
Are you and imbecilic sociopath? Do you want to cause untold mayhem, but you're not smart enough to figure out how? Don't worry fam - we got you covered!
^that's what I brainstormed just now, and I asked 4o to improve it. Here you go:
“Empowering Morons Since 2025”
Tired of watching smarter sociopaths get all the attention?
With LieForge™, you don’t need brains or ethics—just a vague grudge and a WiFi signal.
Upload a face, type a lie, hit "Generate." Boom: You're an unstable god.
Thanks Daniel! I guess we’re maybe all going to stumble on one once :)
Must admit I was surprised when I wrote it - I still wonder if I have read it somewhere, but can’t recall or recognise it. Can’t say I surprise myself often these days. Great article of yours though - it’s an almost unimaginable capability to me. Unfortunately very weaponiseable, I agree.
Yeah, I actually Googled it after your comment to see if it's a quote from some show, but it doesn't look like it. You must trademark the phrase officially, ASAP!
Deepfakes were already relatively easy with lots of off-the-shelf tools, but this appears to make it that much easier.
Sorry, the only acceptable avatar is the man with the stubby beard. In the future, that will be the golden standard for men and women alike. It's called dystopia for a reason.
And yeah, I've watched a couple of documentaries on CRISPR and biohackers. The difference isn't the potential level of threat but the timeline. Deepfakes are easy here and now. If there's something that lets my neighbor clone a biosoldier in his garage tomorrow in a few minutes, I'll start worrying about that.
Also, DeepMind is basically Google now - I believe they bought it all the way back in 2014. Many Google releases we see these days come from the DeepMind division.
Speaking of dystopian forecasts, have you already read the AI 2027 take? You might enjoy it: https://ai-2027.com/
I agree that it seems hard to find a positive use for this kind of thing. What I could envision happening is essentially the end of the video testimonial for product reviews. Nobody will ever believe a "busy mom's" TikTok style review for an egg-whisker, skin cream, hair supplement etc again.
It will eventually make consumers extremely cynical of everything they're ever shown and basically break advertising.
Maybe that's okay?
That'd be the "happy" outcome for sure!
Sadly, I don't think we can count on the average consumer showing that level of discernment.
Case in point: All the Facebook pages sharing clearly AI-generated images of "dream houses" and "my 3-year-old built a giant sand sculpture of Jesus all by himself," and the thousands of likes and comments they get.
Not to mention all the traditional hoaxes, scams, etc. that continue to exist and thrive even in the Internet age.
Thanks! I didn't want to sleep tonight anyway.
(Pithy, but yes, these are even more creepy, and it's gonna be so stupid easy for imbecilic sociopaths to put these to use)
Someone should definitely work "Imbecilic Sociopaths" into the marketing material for future deepfake tools, somehow.
Are you and imbecilic sociopath? Do you want to cause untold mayhem, but you're not smart enough to figure out how? Don't worry fam - we got you covered!
^that's what I brainstormed just now, and I asked 4o to improve it. Here you go:
“Empowering Morons Since 2025”
Tired of watching smarter sociopaths get all the attention?
With LieForge™, you don’t need brains or ethics—just a vague grudge and a WiFi signal.
Upload a face, type a lie, hit "Generate." Boom: You're an unstable god.
Deepfakes. Now for Imbecili Sociopaths! Don't be a square, be a sociopath. Get yours today!
"Don't be a square, be a sociopath" cuts a little too close to home here in the US right now.
I’m the fourth wall, and I’m broken ™️
Thanks Daniel! I guess we’re maybe all going to stumble on one once :)
Must admit I was surprised when I wrote it - I still wonder if I have read it somewhere, but can’t recall or recognise it. Can’t say I surprise myself often these days. Great article of yours though - it’s an almost unimaginable capability to me. Unfortunately very weaponiseable, I agree.
Yeah, I actually Googled it after your comment to see if it's a quote from some show, but it doesn't look like it. You must trademark the phrase officially, ASAP!
Deepfakes were already relatively easy with lots of off-the-shelf tools, but this appears to make it that much easier.
That's a Hollywood movie level of quote-worthiness right there!
I could easily turn all of my essay voice-overs into rendered videos of me reading them and load them on Youtube?
Not with this specific photo-to-video feature: It's limited to 30 seconds per video and is intended for quick messages.
For your voiceover use case, you'd want to go the old-school way with a more professional customer avatar from either Heygen or Synthesia:
https://www.synthesia.io/features/custom-avatar
https://www.heygen.com/avatars/ai-video-avatar
Oh. I did see that limitation on a second review. But I’m sure we aren’t far!
Yeah now that the underlying tech is out there, extending it to longer videos seems trivial.
Sorry, the only acceptable avatar is the man with the stubby beard. In the future, that will be the golden standard for men and women alike. It's called dystopia for a reason.
And yeah, I've watched a couple of documentaries on CRISPR and biohackers. The difference isn't the potential level of threat but the timeline. Deepfakes are easy here and now. If there's something that lets my neighbor clone a biosoldier in his garage tomorrow in a few minutes, I'll start worrying about that.
Also, DeepMind is basically Google now - I believe they bought it all the way back in 2014. Many Google releases we see these days come from the DeepMind division.
Speaking of dystopian forecasts, have you already read the AI 2027 take? You might enjoy it: https://ai-2027.com/