I work and IT, and when I heard the Obi-Wan clip, I immediately thought “ we are fucked “ so many companies use voice recognition to verify that the person on the other end of the phone is who they think it is.
It's not. The "uh"'s come from the original recording of a real human, then the AI just changes the voice to sound more like Jordan Peterson or Joe Rogan.
I watched a presentation years ago where engineers at Google actually showed that they had cracked realistic computer voices (with natural pauses and mistakes) long ago. But what they found was that if people's voice assistants on their smartphones sounded too realistic, customers would get creeped out. Siri and Alexa and whatnot sound robotic *on purpose* for the comfort of the users.
The scary implication is that audio evidence could eventually become more easily disputed (or alternatively manipulated) during an actual court case that would have major consequences on someones life if this technology evolves further.
i mean the audio would still have metadata on where and when it came from however if the government wanted to use this for propaganda against a war enemy or country that would be petrifying
when the ai can take input directly from your brain waves, it can generate an endless stream of audiovisual content thats always exactly what you want. sounds very healthy and wholesome.
I soon as I saw the words "thats always exactly what you want" I immediatley realised how bad this would be. Imagine a pedophile getting access to this technology.
1970 Larry Niven had an idea of the droud. Basically jack into the pleasure center of your brain. Cops find people starved to death next to a power outlet. Doesn't even have to be visuals. Just the parts of the brain that make you feel like you're seeing what you want to see.
If Charlie places all his videos in an AIs database he could deepfake a video, make an AI write a charlie-esque script with his dry humor, and then the audio copying his easy to copy monotone voice.
Jordan Peterson's voice is so unique that the fact that this ai can easily recreate his voice without any problems is absolutely horrific. Like all the stuttering and pauses are so realistic and terrifying
The other way around would be more impressive (or impressive to start with). Of course it will have plosives if the recordings the voice is trained on has them.
@@Sid_Killit’s probably a text generator tied to text-to-speech, as opposed to actually synthesizing the spoken language, but I mean, it sounds like a robot to me.
The breaths are not put there by AI, they are present in the original audio it's fed and just replicated. The technology is speech-to-speech, not text-to-speech, Charlie said it in the video.
It's scary because if there is actual footage of a crime, the person who committed it can just say "that's fake, it was AI generated." If there's no DNA evidence or Alibis, to be found, it becomes really hard to tell what's real.
Nah, you know what's really scary? The tech we currently have is roughly about ten years behind what DARPA and similar departments have been creating and using for the different governments around the world. Makes the last ten years of s*** we've seen and heard HIGHLY suspect.
@@Kushufy I feel that depends on how easy it is to generate a video with a program at home. Plus, it makes it more confusing because it has to be determined if the video was faked in the first place
@@shin-ishikiri-no no... if you are personally recording it, its just an eyewitness testimony with extra steps. if its a security cam then youd need to have a job with access to it and you would also need to be incredibly careful with leaving no physical evidence you uploaded to the server. in both cases youd need to know how to edit the files even in the hardware so that theres no evidence of it being an uploaded video instead of a normal recording, which 99.9% of people on earth do not know how to do. even if the video itself could be perfect, criminals would not be able to do the rest. remember how the person framing would also need to know where the framed person is during the fake recording, if theyre on a phone call, with multiple people, in public around cameras etc the framer is pretty much caught and this is all random problems i see in 5 min, the moment this is a real possibility in the world theres gonna be a ton of people on this, figuring out and inventing more ways to find out. if you are thinking of killing someone, it is going to be way safer to just do it the old fashioned way than trying this complicated super villain shit. just look at that murderer who tried to fake a twitch stream. what you should be afraid of is this tool being used by governments and institutions, not criminals trying to frame you or whatever
@@yahwaysaywhat2998 dude can you shut the hell up kindly ? We get it everything is negative nowadays because of negativity and everything should be negative as usual negative because no happiness exist oh no I’m an ai. Seriously we get it you don’t have to be the cool guy that acts tuff “oh yea we’re in for a wild ride 😎” shut up damn
I'm more worried about false flags and starting wars. Take for example the declassified documents from the Cuban missile crisis. The cia were cooking up some serious false narratives just to rile up the American population. If you haven't read them you should, and then you'll understand just how terrifying it will potentially get.
honestly, this level of AI terrifies me. its just so unsettling how smart AI has become while also just being openly available to everyone on the internet
It’s open to everyone on the internet to keep feeding off of all of you and growing. STOP USING THIS TECHNOLOGY NOW. Stop supporting it stop acting like it’s okay stop letting these people lead us to our doom!
I wanna see an endless episode of Blue's Clues. Where Steve just goes around finding an infinite amount of clues, never to actually sit down in his thinking chair and put them together.
amazon audibles are already in AI voice if they don't have a person to voice them. i would bet they would be the first ones to steal business away from authors by just replicating their voice.
@@baileykerkvliet2588 Is it an AI replicating the author's voice or is it the author replicating an AI replicating the author's voice? We'll never know. Society will never be the same again.
Defending yourself in court in 10-20 yrs is gonna get a helluva lot more complicated, when the prosecutor turns up a fake video or picture of you walking away from a crime scene.... I'm more afraid of that then Skynet tbh
"who doesn't giggle when listening to obi-wan passionately recalling his adventures with the rise of the sky cocker" another incredible passing line from Carlie over here
Thinking about legal ramification for evidences in trial, the company who produces these AI recreation probably should encrypt some sort of code withing the Audio/Video file they produce to distinguish it as AI generated
@Wutai Nation | Entertainment & Video Games im certain at least in theory you can code the AI to give off a near imperceptible noise that can distinguish it as being AI generated while also allowing it to sound real to anyone listening.
if there are people on the side of the law there will be people against it, it's the rule of life. even if they do encrypt 100% there will be some companies that will somehow bypass this or even single coders that will make their own AI without encryption and spread around. no matter how hard you try, you can't stop piracy
Something that comes to mind is how the Terminator was able to perfectly imitate John Connor's voice. Interestingly enough, we're now living in the decade they hail from.
And there were alot of people who laughed at the concept of Skynet and the rise of genocidal machines ever happening back in 1992. This is terrifying to see happen right before my eyes. We are only a small mis-step away from a dystopia of a world being ruled by robots. God, I hope scientists are careful with gene therapy or manipulating DNA in the future, otherwise we might be give birth to humans with mutations and powers like in X-men. Humanity truly is it's own worse enemy 😬
As someone who's worked a while in AI, I'd say that this perceived growth of AI is moreso due to the development of new web platforms to host these algorithms. There's a lot of groundbreaking work that's been done behind-the-scenes for the past few decades, particularly within the past 15 years. This technology has only become more widely available, likely due to cheaper computation and server costs.
I'm also into it. One thing to note about the language and speech models is that they are not properly global. Out of Western languages I haven't seen any good emulators besides English. Loads of work needs to be done to feed multilingual data and connect this data to understanding (to my understanding the text-to-speech works properly only if the context is understood). Though I will admit ChatGPT is quite good at translating even smaller Scandinavian languages. And yes, you are correct... Loads of servers and digitalised data is needed - not saying it in a bad mission impossible kind of way - just saying.
I also think that it is getting overhyped. For example, claims that AI can now fully replace software developers are way off. The amount and scope of cognitive tasks required to fully develop software from the ground up will remain out of reach for AI for a long time. Instead, I expect AI to become an assistant. There are many cases where a software developer has to write some boilerplate code, a task that is both tedious and time consuming. One example would be a function that takes date and time columns from a CSV, converts them to one combined ISO 8601 datetime column, and replaces the original two columns with this one. That is annoying to write, because it is usually something that comes up as part of a much bigger task, and _that_ task is what the sw developer is actually focused on. But with an AI, the dev can tell the AI to write that function. The dev then reviews the generated code (which is urgently necessary because the AI can be very wrong sometimes), integrates it, boom, done.
You godda love humans. Instead of using A.I. to automate the shit we don’t like so we have time for art and music, we INSTEAD use it to automate art and music so we only have time for the shit we don’t like!
Joe Rogan voice must be good with AI because of how many hours or material you can get from the podcast, the more you can feed the AI better they can be
@@DoctorWu23 lol your phone is always listening and watching you. Plus we probably in simulation anyway. We will be creating numerous simulations soon, then those simulations will create simulations. We don’t have free will. So everything that is happening is predetermined and with law of physics there nothing we can do about it. We just watching it or thinking we are from our own perspective.
@@DoctorWu23i wonder how much news we’ve seen over the years is fake. it’s funny alex jones said SH was fake, a video showed a crisis actor that was a “dad” of a victim laughing right before getting emotional for camera. months later and we see exactly how what he said could’ve been true. we see images and audio that look realistic asl.
THIS is actually super deadly like you said moist, you can legit fram and sue someone by saying things from just that website its too deadly it can ruin careers lives!
I'm more worried about false flags and starting wars. If you haven't read the declassified documents on the Cuban missile crisis. The cia were cooking up some serious false narratives to rile up the American population. This will make the stuff they planned to do soooo much easier.
The best AI joke I've seen so far was Larry going to a doctor and the doctor asked if he wanted the good news or the bad news first. Larry said "Why not both at the same time?". The Doctor: "Ok, you're pregnant" 😂
I was having a conversation just last week about using something like this to put your own voice in a video game as the main character for full immersion
@@MrFrigid247 I think this is going to create "The Ultimate Game". Like, you log in, and tell it exactly what you want to play (Shooter, RPG, etc.) and it's going to generate an entire goddamn game for you with voiced characters, maps, and quest lines. And the next day, you can play a totally different game. And the day after that, and after that. Or video games can be literally never-ending. It'll keep on generating content as long as you want to play. Finish this main quest line? Great, here's another for you to play that's totally unique. So long that you still like the gameplay, you can keep on playing it forever and never see the same mission or storyline twice. I predict within 2-3 years, we're going to see a game like this with generative AI that does more than just generating maps, spawn loots, or control basic, pre-scripted NPCs.
You know, John Carmack after spending his time working on advancing VR, said that his next big quest was Artificial Intelligence. The timing of this advancement in tech could not be more exquisite.
This is inevitable, whether we like it or not, the only way we can make count measures for it is if we change our shoot first ask questions latter policy for drama.
Should theoretically be easy (relatively speaking) to start training AIs to recognize AI created works. Then it's kinda a software arms race somewhat like what we already have, if on a higher stakes level. Then, I'm a layman so what do I know?
If this is available to the public, we have to assume this tech has been around for some time, which potentially calls into question several instances of recorded voices in recent years.
I would love it if society fell apart because people were afraid that other people were going to create a robot to call them gay in their own voice…. i’d love to look back while I’m eating my neighbors forearm and chuckle
Yeah, the Gandalf deepfake skit is halfway funny. Sir Ian lounging around in a dressing gown and slippers, slagging off the Orcs whilst drinking tea. The gym pullups scene is worthy- " fly, you fools!" 🤣
@@dereksegra7738 That's so true!!! I've watched shows with my boyfriend (who is fluent in Chinese) and there was one character who was supposed to know Chinese and be fluent and when he spoke it, it was complete gibberish
Kinda terrifying tbh. If someone you are no longer friends with/an ex hates you and wants revenge they could feed it a few voicemails or audio of you talking, make you say you’re gonna go do something violent or confessing to something horrific… I feel like it shouldn’t be this accessible. Hopefully there is a way, if not I hope soon there will be a way to absolutely confirm or deny things like this.
There absolutely never will be an easy way to confirm or deny. Absolutely never. Sucks. The claims are no longer evidence because they can be extremely easily faked.
AI generated arts have a type of invisible watermark we can't see but is there and can be identified by computer so maybe there will be a similar way to identify ai generated voice too
The worse part than the bad actors is that as you’ve mentioned, even normal evidence can be thrown out and we could probably see an uptick of crimes because of this. I’m not sure how prosecution will move forward after this
Yep, bad actors will use it both as a weapon against innocents and as self defense for crimes. It basically creates a situation where we cannot trust any recorded information, which is bad.
This is basically what I thought was going to happen tbh. Everyone has been freaking out for the past 10-20 years about pervasive surveillance, photos and videos and violations of privacy, and I was just like "Ok so camera drones will become small enough to be indistinguishable from a large mosquito, but AI is going to make this stuff worthless before that happens" and here we are. Now it doesn't matter what people record you or photograph you doing, because nobody will ever be able to tell if it's real or not.
Another AI may be able to, but the problem remains if human validation is impossible, whose to say if the validating AI is mistaken or tampered with in one direction or another?
Aw heck guys, that whole thing about "Nobody will be able to tell if it's real or not". I've been thinking about that as the AI gets more and more powerful. What happens when this AI starts being used on Politicians, Governments and so on.
surveillance is still completely possible cause the government and whatever other actor needs surveillance still trusts their own systems. This AI generated stuff is only a problem in a court of law where everyone is trying to establish responsibilities and damages and so on.
It's crazy how in just a couple of years, AI has been developed to create realistic deep fake images, video, voices, simulating fictional characters and being able to answer just about any question possible.
Late last night I watched the AI Seinfeld discussed whether or not it was possible for a human soul to be placed in a tree and retain cognitive thought. In the end the gang decided they would try to find a wizard and ask him. That is exactly the kind of thing you should ask a wizard. The AI managed to connect souls living in trees to magic. It's getting scary good.
Imagining a world where every creative job is done by AI is like pretty horrible. Back in the day people "only" got robbed of hard labor jobs. Now they get robbed of the jobs they love. Awesome. Can't wait for the sequel
AI will inevitably only be able to "create" based off of what already exists. It's very difficult to get AI to the point that it creates completely on its own, independent of any previous input. It isn't taking creative jobs, it's just advancing it. Edit: No shit humans do that, I thought that would be an obvious one. The difference is that you can lock a human in a room without ever seeing the world and it'd be able to create art, the AI wouldn't. Jesus, people are so dumb.
@@vloppysagina That sounds like what humans does. We also take input and creates something creative based on that. Try and think of something that does not relate to anything else. It's impossible.
Good, you shouldn’t be paid a living wage just for drawing some pictures or doing something creative. most people work really hard to make the money that they have. I don’t feel bad for you Freelance artists that make tons of money by basically doing stuff that you would do it for free anyways. The fact that you make money by doing fun freelance work that doesn’t require much work Other than just being creative is a gift/privilege. I feel bad for people who work at McDonald’s and have to bust a butt all day for minimum wage for whatever reason nobody cares about those people who work so hard to provide a service every day to people instead, they care about these low life, streamers and artists that work two hours a day and complain that they are losing their job. It’s about time they get a real job.
Why would you want to do this? Seems like a waste of tech. Much more important are being able to somewhat speak to NPCs properly, and possibly having entirely AI-driven games, where AI can create stuff like quests and events on the fly, including taking into account what the player already done, vs. everything being static and pre-programmed.
too bad that discussion has already been had. we the common folk don't have the privilege of being involved. If the public has this access think about what the public doesn't have access to.
The reason it is so good at Joe Rogan might be that his podcast could have been used in the training of the base model. The way that theses models work, is usually that you "pre-train" them on a lot of different data, and the you "finetune" it on a much smaller set of data, to get it good for one specific person. Joes podcast would make it very easy to generate a lot of data for the base training process.
This just proves more and more that Veggietales not only predicted randomly-generated comedy, it was a legitimate prophecy of robots becoming comedians.
I’m calling it now. Voice acting in video games is going to turn completely into the utilization of ai voices of famous people. Pay them a little bit to use their voice without having to pay them the big money to actually have them come read lines. You could get so many big names by doing that. You could honestly probably use this for animated tv shows too. And obviously to keep using peoples voice after they die. Voice acting in general might die
@@IndyJacksonTT Exactly this. The prosecution actually tried to use this in the Rittenhouse trial. The program can't actually clear up an image, it just adds pixels that it thinks would go there. It's essentially a computer generated image and should not be used in court as evidence.
@@FUNKOfilms It can be used like DNA can be used, it’s not 100% but you can sell it to a jury that it’s 99% what “would have been” there. It’s up to the jury to believe it or not
Watched it when there was only a few hundred viewers and it kept trying to pull off the joke “what do you call a bear with no teeth” but it couldn’t finish it. Everyone lost their sh** last night when he immediately yelled “gummi bear” right after and proceeded to nail out 5 jokes back to back. It was amazing.
That Rogan/Peterson conversation was insane. If Adult Swim uses this type of AI for Rick and Morty, they won't need to pay for anyone to replace Roiland.
@@BiggerBossN313 yea that would be the next logical step but they need a sample of voices to take from originally in order to make the AI voice in the first place. maybe they can have it so they give the human VA the option to either sell their voice to a company for a fixed fee or do the voice themselves for the same amount of money
I’m now scared that every Charlie video has been an ai generated one for the last week or so, and in a few days he’ll come out like ‘you STUPID idiots’ I’ll cry 😭
Human creativity is filled with repeating patterns. Art is filled with math (golden ratio, symmetry, perspective, etc.), books are filled with common tropes that work, so are movies, characters have common looks tied to their traits, etc. Up close these patterns are not clearly visible, especially when each piece of media is still a lot different and unique from another. But feed to a computer a large enough sample and those patterns become visible, it can learn them and build upon them. And to us the results it will produce will seem good, because its using common patterns we already discovered to work.
@@foppypoof5195 What sucks is of all the future dystopian realities, Battlestar Galactica is the lamest. Really gonna be out here tryna guess who's a Cylon when we coulda been worried about accidentally becoming our own grandfathers, lol
There's actually a technique called generative adversarial networks, which makes two AIs compete one to generate something, the other to tell if it's a fake or not, and you get a feedback loop of improvement :D
It usually goes like that, currently technology will pushed to it's absolute limits and will get to a bit of a standstill, something new gets developed that suddenly pushes everything we knew that could be done, technology goes on a bit of a drug-induced marathon and goes a bit insane for a while, new steps taken everyday, and then we reach the standstill again, doing everything we could with what we made. It just makes me wonder when we'll reach the true limits, if there even is one
My favorite Nothing Forever joke I've seen is where they're talking about how they saw someone advertising a device that can "create meals in five minutes". They proceed to lampoon the idea while one of them is using the microwave in the background.
Dude, yesterday I was watching that AI Vtuber, Neuro-sama, and she pulled up (unprompted) the Kurzgezagt video about automation and robots advancing enough to replace humans in different jobs followed shortly by two other videos that both featured mention of robotic arms and I was quietly trying to not have a small panic attack about whether that was just coincidence or not for a while after that. AI tech has gotten VERY cool and VERY advanced in many ways recently but "Pandora's Box" really does feel accurate for describing it.
AI Vtuber... this is a thing? I was just wondering how long it would take before someone used an AI to run a Vtuber that could stream literally forever.
Yeah, it’s been developing crazy fast I thought it was just chatgpt and than characterAI and the voice thing started picking up traction and now that shit is scaring me.
@@natenotsogreat1054 Change is always scary. I remember the personal computer scare in the 80's, and when video games "ruined childhood", and all that crap. Adapting is not fun, and people don't like to adapt to new things. The creepiness factor is almost always a big fizzle, after the fact.
Can't wait for the Dune-style holy crusade against AI to begin lmao. It's wild that we're living through an era where AI is becoming something we have to consider in our daily lives. A few days ago I had an exam for nuclear physics and because the course wasn' t taught very well me and friend decided to meme it up by asking chat gpt questions. I swear its answers were more structured than my professor's teaching... They were wrong , but they were easy to understand.
Well is that really AI or is that just AI taking examples of well structured answered and filling in blanks? Because AI as of right now is just one big randomization button with the entire internet as a template. Not really thinking, just computing based off of our prompts. AI isn't necessarily scary because we as humans don't have the tech (assuming it's possibly) to quantify the human brain. But the COMPANIES that have free and total access to ALL of this data all these ai are taking from people is fucking insane though and no one is talking about that much more realistic and upcoming threat. It's always been about the companies and how they use their products vs the products themselves
That Dune thing is so fucking funny to me We, as a species, know about the dangers that technology can poses. We knew about the danger of climate change for around 100 years (back when the signs of it were rather miniscule but still visible) and we did nothing about it. Half of the sci-fi universes touch on the danger of AI and we are ignoring it cuz some tech bros and corporations said "nah it be fine" Sometimes I think that we actually deserve to go extinct
@@sono_chi_no_sodium_chlorid7635 That's because our "philosophical level" on average is too low as a whole and what is more important imo, is that technological level is speeding too fast, so the only hope is to government to impose it, but they kinda don't care also or scared, that things they need to do is hard for us to understand, like how children don't understand why too much candy is bad And let's be real, there is no need for 7bil people on earth, so all this "all art work is dead" or climate change will just help to impose the old gold rule of "survival of the fittest"
For the AI seinfeld, they should add a feature that when it tells a joke, the chat can vote on whether it was good or not, so that the AI can learn what humans think are funny/not funny
See the problem with this is this is exactly why facebook (?) did or some other big company; and it derails pretty fast as edgelords make it appropriate racism and shit.
That joe rogan one scared the shit out of me. You can BARELY, BARELY tell it’s not real. Only way to do it is to listen to the beginning and end of each word and hear the tiny tiny robot like transition.
The original actor for obi-Wan Kenobi was a man named Alec Guinness. And he passed away 23 years ago. It’s almost unbelievable how this Ai got his speaking pattern down. Man, it’s scary thinking about how people can abuse this program.
Its all so strange. 10 years ago, people imagined that in a world made of robots and AI, it would be humanity's creativity that would set us apart. Everyday, it becomes more obvious that we were completely wrong.
Yes its not human creativity but theres something about a human a real connection that makes it more interesting. We may be worse in it but its authentic like music etc
@@theresurrection33 I predict a post-work world within 20 years, and my friends all argue that human connection will always be financially viable. What'll actually happen though is that we'll get AI to make us Bladerunner-type androids that are indistinguishable from real people.
I completely disagree. These AIs still require prompts and at the end of the day the ones judging them are humans. We decide what is « good » or « bad » so ultimately we are still in the choice.
Damn man I'm torn cause this technology can be used for such amazing and hilarious things but while also being used for purposes that are horrendous and amoral at the same time, but thats sadly how most things are in life.
for some of my university classes, the professors literally have on the report outlines that we're not allowed to use ai to generate the report. crazy how much this has developed in such a short time
If parents want their children to be a healthy, intelligent individuals, they should never buy them phones or laptops. Maybe consoles at best and family pc that has filters. Sadly this is what it all has to come to. But internet is getting scary.
