Joakim Sorthe
Steve Troughton-SmithHere’s your AI astonishment/nightmare fuel for today:
"TL;DR: single portrait photo + speech audio = hyper-realistic talking face video with precise lip-audio sync, lifelike facial behavior, and naturalistic head movements, generated in real time.”
Mark Jardine 
@stroughtonsmith That Mona Lisa demo... 😳😅
@mark out of nowhere 😂
Dane Deasy@stroughtonsmith this is awful. It should be illegal to create software to do this.
Johannes@danedeasy @stroughtonsmith this is really horrible to see, you are right.
How can any serious developer/scientist work in such a shit?
PS: The internet will be flooded with such easy to produce fake videos, this will end terrible for all of us.
MadeInDex 📰🌎@herbsti72 @danedeasy @stroughtonsmith
#microsoft gives financial incentive to do evil by appealing to peoples need for safety & lower instincts like greed
@herbsti72 @danedeasy @stroughtonsmith
The good thing i see coming from Ai-generated videos is the bankruptcy of Hollywood.
#ChatGpt will spit out a #script customized to each viewer and the #video platform will create an individualized #movie.
I don´t think Hollywood would be the big issue.
I think this technology will be used to destroy trust in media, people and democratic structures.
Which such technology it is quite easy to flood the internet with much more fake videos about politicians and other people to discredit them.
@danedeasy @stroughtonsmith
@herbsti72 @danedeasy @stroughtonsmith
Yeah that's the bad side, the trust in anything not by an old, "trusted" source like a newspaper or popular influencer will be shattered, which might lead to an even further centralization of "truth" online.
Yeah that's the bad side, the trust in anything not by an old, "trusted" source like a newspaper or popular influencer will be shattered, which might lead to an even further centralization of "truth" online.
anmey@herbsti72 ich frage mich ob das nicht eher gut ist 😀 in meinen Augen sind all diese Dinge schon lange kaputt. Jetzt sind wir gezwungen uns zu überlegen wie wir das reparieren.
@madeindex @danedeasy @stroughtonsmith
@madeindex @danedeasy @stroughtonsmith
@anmey @madeindex @danedeasy @stroughtonsmith Ziel von vielen (extreme Rechte, Russland, China) ist es die westlichen Demokratien zu schwächen/zu zerstören.
Diese Möglichkeit so einfache Deep-Fake-Videos zu erstellen kommt da für gewisse Kreise wie gerufen.
Und nein, da wird es dann nichts mehr zum Reparieren geben.
Ich fürchte es wird fatal enden... :(
@herbsti72 unsere westlichen Demokratien schaden sich durch Doppelmoral und durch Kapitalkorruption schon auch genug selbst.. jetzt fehlt nur noch, das das Netzvertrauen durch Quantencomputer komplett ausgehebelt wird.. interessantes soziales Experiment.. ja wir müssten dringend Wege finden wie wir als Menschheit grundsätzlich vertrauen können... fatal
@madeindex @danedeasy @stroughtonsmith
@madeindex @danedeasy @stroughtonsmith
Jim Bliss@madeindex @herbsti72 @danedeasy @stroughtonsmith
You're seriously suggesting it's a good thing to replace the artists and crafts-people who work in film with an AI that rips ideas from existing artists? You want to destroy the careers of writers, set designers, costume designers, directors, actors, key grips, carpenters and the other hundreds of professions employed in film production?
Why is that a good thing exactly? And do you think it stops with Hollywood? Seems very short-sighted to me.
I think it's 2 debates:
1. The environment
2. Art / Creativity & The result for the consumer
1. I think from an environmental point of view AI could be good in certain aspects such as:
smart & creative people allocating their time & skills towards tasks which are more related to solving problems which might help the environment.
And a reduction of resources used to produce things such as entertainment.
2. Very philosophical, i think all arguments here make equal sense.
@JimBliss @[email protected]
2. A lot of answers exist for questions such as:
A. Is there a difference for people if #art is made by #AI, if they can't tell the difference.
B. Does humanity still need art even though there is more art than can be consumed in a lifetime?
C. Is it important for people to make art as a full-time occupation & would there still be progress in art if that happens?
D. Will AI be able to create original art & new art direction?
And many more
2. A lot of answers exist for questions such as:
A. Is there a difference for people if #art is made by #AI, if they can't tell the difference.
B. Does humanity still need art even though there is more art than can be consumed in a lifetime?
C. Is it important for people to make art as a full-time occupation & would there still be progress in art if that happens?
D. Will AI be able to create original art & new art direction?
And many more
I think AI progress is very scary (it can already do most of the things i spent my professional life on learning better than me) on one side and endless opportunity on the other.
I think fear is a normal reaction to many technological breakthroughs (even though this one might be different), especially as we are not able to control it.
nutbunnies@stroughtonsmith Continuing to wonder how the good outweighs the bad with this kind of thing
QuarterSwede@nutbunnies @stroughtonsmith It doesn’t. We’re already being lied to, now everything is going to be lies and we will literally have zero idea if anything is real or not. What’s the point if there is zero reason to trust anything. This is extremely bad for society.
Matthew MillerOMG. "Nothing that could be problematic here! Moving along...."
Laurens 🐐@mattdm @stroughtonsmith The "women" have very similar eyebrows, eyes, noses and mouths, all just a tiny bit imperfect - but all in the same way. That pattern is still too visible.
Kathy E. Gill 🇺🇦Dear lord.
KanaMaunaSo what will be the main use for all this, beyond tricking the gullible?
Radieschen
Enhance educational equity - Student A from a rich district gets a real live teacher and student B from a poor district gets a mindwarping monstrosity.
sheislaurence@radieschen @KanaMauna @stroughtonsmith thank you so much for sharing the supposed drivers behind this development. And the way it is written is revealing “potential benefits”: like the purpose is indeed societal destruction, disinformation, creative theft and ultra capitalisation, they came up with last minute bullshit pretexts like “companionship”, seriously?!! Show me the research showing the mental health benefits of talking to a moneymaking pixel wall!
I guess the purpose is just making money but accepting that people will do bad things with it.
When I read the beneficial uses, I was wondering whether they're actually needs by the groups they're mentioning or if it's just privileged people thinking this might be helpful use cases.
@radieschen @KanaMauna @stroughtonsmith the latter, in fact I really believe it is more cynical than this, clearly there is zero market research about the applications they mention, and no one who loves their money go into business without market research. So the research they did and know is lucrative is selling to psychopaths, dictactors and arseholes.
@sheislaurence @radieschen @stroughtonsmith
Exactly. HOW will this be used in education. It is such an afterthought.
@KanaMauna @stroughtonsmith
AI will make custom movies & tv shows & games -> bancruptcy for the entertainment industry -> might be good for the environment if all these smart & educated people do different jobs instead
AI will make custom movies & tv shows & games -> bancruptcy for the entertainment industry -> might be good for the environment if all these smart & educated people do different jobs instead
@madeindex @KanaMauna @stroughtonsmith
Down with artists!! Down with crafts-people!!
Let's fill the airwaves, the pages of our books, and the walls of our galleries with AI-generated content stolen from artists and crafts-people instead.
Down with artists!! Tell them to get "proper jobs" and leave the creativity to the executives who run AI companies.
You can't be serious.
@JimBliss @KanaMauna @stroughtonsmith @nlpbot
People will always make art i personally think, but maybe not as a full-time occupation anymore. Depends how good #AI gets at it. I think it is already much better than me at many things i did as work (i don't like that it's often assumed online that somebody is not affected themselves by something they talk about). There are open source alternatives & people can train their own AI, it does not need to end with big corporations owning it.
People will always make art i personally think, but maybe not as a full-time occupation anymore. Depends how good #AI gets at it. I think it is already much better than me at many things i did as work (i don't like that it's often assumed online that somebody is not affected themselves by something they talk about). There are open source alternatives & people can train their own AI, it does not need to end with big corporations owning it.
Tytrater@madeindex @JimBliss @KanaMauna @stroughtonsmith yeah I don’t think artists need to find new jobs, I think they should be considered partial owners of the models that use their art. Artists should get a royalty for every piece of their work that’s fed into a gen AI model, as well as shares in the companies that build and deploy gen AI, that way they can focus on their art and even let the AI worry about monetization if they want
@Tytrater @JimBliss @KanaMauna @stroughtonsmith
Yeah that would be great if artists were compensated for AI using their work to train. I could imagine it similar to the way music is paid on a per-use basis. There will probably be some interesting proposals in the near future.
Yeah that would be great if artists were compensated for AI using their work to train. I could imagine it similar to the way music is paid on a per-use basis. There will probably be some interesting proposals in the near future.
@madeindex @KanaMauna @stroughtonsmith
We'll agree to disagree.
I can see useful applications for "AI" (assuming we can deal with the horrendous power usage) but the creation of art is not one of them.
In fact, the idea horrifies me and I suspect will lead to a very dark future if it gets widespread adoption.
But I've already been called "scared and stupid" on this thread. I thought I'd left that bullshit behind on Musk's hellsite. So I'm stepping away from this discussion if that's OK.
Brian Grinter
Scott Jann@stroughtonsmith I made pretty much the same thing on classic Mac OS using MacInTalk to make a cartoon face animate like 30 years ago https://www.zenwheel.com/software/mouth.html
broti@stroughtonsmith what the… nightmare fuel sure did it’s job.. 🫠
xs4me2We are doomed. There need to be worldwide rules and regulations on how to apply this technology soon or this will go terribly wrong.
The EU now has the digital services act…
Christer
@stroughtonsmith It's interesting, #AI has gotten to the point where you can only trust information coming from a source with a great and long record.
So the intelligence community just needs to spam massive amounts of fake documents & other data & a whistleblower can never be trusted again, without face to face verification 🤔?
- That person could always have been agent with fake information, but now it could become almost impossible to find the real ones right?
Sumukha S@stroughtonsmith Deeply concerning.
Bad News Nobody@stroughtonsmith Fortunately it's still possible to spot the fake on a big enough screen (look at the teeth for example), but i'm afraid it won't be for long.
lippyduck@stroughtonsmith At the very least, apps like this should be obliged to insert a digital watermark stating its provenance and social media sites should have to extract and display it.
🇸🇭🇮
@stroughtonsmith I get that this is worrying, but "hyper-realistic"? this woman is growing teeth in front of your eyes as she speaks 😅 (and that's not even getting into the quantum tangled hair that sticks in place as the head moves)
@shi @stroughtonsmith Right now, yes. But as I said elsewhere, if you are viewing it low-res on a social media video on a small screen and only giving it half your attention then some people will fall for it. Especially people who don't realise it's AI and are not looking for telltale signs.
Then in 5 years it'll fool you, me and everyone else even on a high-res IMAX screen. This is a nascent technology that is already massively more convincing than it was 12 months ago.
@JimBliss I viewed the video on a low-res screen on social media 🤷♀️
also I think you'll notice I didn't comment on possible future developments, just calling the video uploaded above "hyper-realistic" - when (even at low-resolution viewed on social media) it looks like someone who did a weekend course on Live2D tried animating a still photo (but some gremlins popped by to fuck with the teeth in the animation)
@shi It looks far more realistic to me than someone who did a weekend course on Live2D.
Also, you watched it with prior knowledge that it is AI generated. My point is for those who are doomscrolling through TikTok and not applying their critical faculties.
And I know you didn't comment on possible future developments.
But I did.
And the reason I did was to point out that we should be looking at ways to deal with this kind of thing while the telltale signs are still present.
asol@stroughtonsmith This is amazing! Imagine how this can revolutionise customer service.
I love this example so much 😂
I love this example so much 😂
Steve Riggins@asol @stroughtonsmith you mean when the customers are also AI, because that’s how it all ends. Westworld. lol
@steveriggins @stroughtonsmith an agent to talk to an agent, agentception
Nini@stroughtonsmith Ugh, what is the goal here? To make us so unsure about what we see, hear and read that we trust nothing?
Marco Ponton@stroughtonsmith Today we’re trying to protect our parents from telephone scams. Tomorrow our children will have to protect us from FaceTime scammers. This is both fascinating and quite scary.
S★m V★rm★@stroughtonsmith thanks for that
Tamir Bahar@stroughtonsmith so much of my brain power is going to be focused on "did their teeth just change in size? What happened to that ear? Hair shouldn't merge as ot moves!" now...
erin sparling@stroughtonsmith @patters this, but for hedcuts like my profile image.
Jarrod Blundy@stroughtonsmith Just wild.
@stroughtonsmith “it could still potentially be misused for impersonating humans”
Hold on while I find Sherlock
John Ellenich@stroughtonsmith these kind of look like higher resolution/more detailed spatial personas… getting similar vibes.