People don't realize how dangerous this is. Anyone can take pictures of you and use that other AI to copy your voice and make you say whatever they want. This is chilling
@@guimcast1 in 10-20 years they will probably be completely indistinguishable from a real video. Instead of developing software for identifying fake videos, we'll probably have to go the opposite route and validate real videos through a token system or something like that
I also noted that the way he said A.i is a bit weird, but i had no reason to believe the voice over was fake. I simply scoffed it up to the British accent making A.i sound weird. I'm a bit shocked.
I'm excited and terrified because this shit is coming fast a furious now. Machine learning and AI-generated everything. Just like they said it would, just like we all knew it would, but now that it's here and the Boston Dynamic robots can run around like people, and the humans can't tell if they're talking to a person or a computer and even my PHONE - an iPHONE 7 is showing much greater levels of awareness than it did a few years ago (same phone, updated OS). AI-generated images are fucking insane now. A few year's ago, they were not even abstract art, they were just nonsense vomit. Literally two years ago. now, they are photorealistic. Within next ten years, we'lll have full length movies that are 100% generated by AI. But real intelligence all around, genreal intelligence will still only exist in a few super computers. Those super computers are borderline sentient, though. That "jedi" joke by the Google super computer AI is funny, witty, clever, and clearly a level of understanding that exceeds that of my five year old.
Man, even your voiceover had me convinced it was being delivered by a real person! I didn't catch on that it was AI until you had to pronounce "debatable" haha
This actually scared me. I thought the voice was real. I heard the breathing and was totally convinced. Took me until 1:18 when he didn't pronounce the I in Deepbrain AI. Although I should've noticed when he immediately mispronounced Simon Cowell.
Thats why AI will feed us "truth" it deems necesery, total monopoly over information, you cant use your "critical thinking" if objective facts are indestinguishable from deepfakes, Metal Gear Solid 2 predicted future 20 years ago
If you know what you're looking/listening for, you can still spot AI generated content. Examples: 1:16 Deepbrain Ayy 1:58 De-bat-able 4:29 High perr-ealistic 5:15 Sensity is creator~ platform 6:04 A similar initiative _bylays_
@@exonet51 "far" is relative. These pronunciation errors will be fixed in a few years and the tell tale signs will become ever more subtle to the point where the vast majority of people won't pick them up!
We already live in a world where half the people can’t agree on objective reality. With this technology, within a few years, there will be no consensus whatsoever on reality anymore. I shudder to think of the ramifications.
No one has ever known reality. But when people don't know who's "real", we're fucked... And good luck convincing the "not" real ones. Fortunately, they're not in the "real" world yet... Okay, I was going to sleep and now I've generated an existential cricis. YES!
@@trulymental7651 but if you do both you are not “ignoring tech” so get off youtube and go read some books, those who say those things, shouldn’t be on any social media to begin with, but just “ignoring technology”
It's the only thing which matters, actually. This is why Huge Companies are now dominating Ecommerce. The public is paranoid, using anybody they have never heard of online. So Google, eBay, Amazon, Walmart dominate sales. Building any of these is not hard, but without the Brand it's pointless.
I dread the day where someone has evidence of someone doing something and that person doesn't get in trouble because they can't "prove it wasn't a deepfake".
by that point no one will take video seriously anymore anyway .... because of this ...... after all it`s already a century old technology and was bound to be compromised eventually
@@EyelessEntity The problem with that is that even metadata can be tampered with. Essentially we will be getting to the point, as was pointed out, that only another AI will be able to distinguish between real and created by another AI, which ironically would provide the algorithms for the fakes to be made even more realistic. This is a snowball that has just begun to roll downhill.
There are a few moments where the intonation either drops or raises weirdly, but you could easily write this off as a poorly edited cut in the audio. This was terrifyingly convincing
When you watch a TH-cam video, you are ready for those short "cuts" and "dents" in recording, as creators don't need to use the smoothest of the most refined film montage. That's the main factor allowing most AI talk failures to pass under the radar.
Especially scary since the majority of people aren’t as aware of the nuances of what makes something detectable as real vs fake. Most people can’t even tell when something is OBVIOUSLY cgi or photoshopped. Most people don’t pay attention or see reality, they see what they predict they’re seeing.
And what's even scarier is that some people WANT to believe false information about people they don't like and won't be convinced it's fake even when presented with overwhelming evidence...
Yeah, I dread the day where someone has evidence of someone doing something and that person doesn't get in trouble because they can't "prove it wasn't a deepfake".
Been looking for a channel like this for a while... non-sensationalist and succinct info on AI updates, and potential scenarios, without jargon and without cheap thrills. Keep it up.
This channel still has clickbait thumbnail though, not that there's big problem with that, but still. Especially this video where thumbnail image never appeared in the whole video itself (1 frame lasts like... 1/30th second does not count, of course)
@Patrick Silva There's been good and bad things going on simultaneously since the dawn of time. It's an eternal struggle. We can only support the good guys, hope for the best and be ready for the worst.
What is terrifying me the most is that the database of almost everyone using social media is already made to be used.. any of our close friends or family is a target if the tech easily reaches hands..
Hence why I came off all social media 7 years ago…. I saw this as threat then and it’s become very apparent now that this technology opens a bigger worldwide problem.
Nah I disagree, detection tools that work like anti-virus tools will most likely improve and develop faster than deep fake technology. Look at what happened to viruses on PC’s. Viruses were rampant in the early 2000’s. But as big money was being pumped into research for anti-virus technology over the years, viruses are not that common anymore. With Bidefender along with taking other easy security measures your pretty safe, and theres still better anti-virus software than that that you can buy.
When I was young I used to think robots would take over all of the laborious work leaving humans to explore the arts and science. Now that I’m older I I’ve come to know that the robots are here to entertain us and I still have to mow my own lawn.
Brings a whole new light to the old saying that taking a photo of someone was stealing their soul. Now, in the wrong hands, it could easily be just that.
Islam discouraged people to take photos of living beings even if it is a selfie . And any drawing or picture of any creature with eyes are considered a living object . Its also discouraged. As the time goes through history , now I am understanding the wisdom of it . Alhamdulillah
I believe you are right on the button Laura..🎯. Somehow I have this odd feeling that the universe is Teeming with life, & at the same time its a graveyard littered with the evidence of species that innovated themselves right out of existance.
This is the type of stuff that was considered science fiction not that long ago. There's still an uncanny valley but it's getting scary close. I'm reminded of the Michael Crichton novel Rising Sun which had video being altered to frame someone else for a murder caught on video. It's a cool tool and when used creatively can do wonders for film and entertainment. But in the wrong hands, this could cause some serious trouble.
That's what more people should be focusing on...what this will be used for in the wrong hands. Or, perhaps, it was always intended for nefarious purposes and things like movies are just the cover.
There's a lot of stuff we've created that is actually amazing. Look at the internet as well as social media. Great tools, yet often greatly abused. Deepfakes are just another thing, amazing technology, but what are the consequences of using it.
This is a very concerning issue. I think that we should have some type of international law that protects likeness and voice prints. Also, if a voice print is used, the original voice actor should be compensated. The could be a very rewarding process for many if it is done ethically.
Not true. This tech will be instrumental in creating digital realities. The downsides are literally trivial by comparison, especially as they are not actually that hard to deal with. Why everyone acts like a fake speech by a president will start WW3, I don't know.
@rasmasyean no, but a lot of people are stupid anymore and do believe everything they see and hear. Kinda like you. Someone who assumes and jumps all in thinking they know everything.
One way to counter this would be. You would be forced to have your own website to post an "Official" release. New outlets would need to embed a link from your source, any other would be considered unreliable.
Yeah... We just had nations lie about how a lab made virus was released... Nations and corporations won't adhere to "best use" ethics. They'll push us all to war in order to further advance this tech, along with AI, brain implants, human augmentations, and social credit scores. This is not heading in a good direction for humanity... Not one bit!
@@katie7748 I think you're on to something. I can't imagine why people would even create deceptive tech like this unless they planned to do terrible things with it right from the beginning.
With the rise of deepfake and AI generated images, actors of the future won't have to look a particular way to fill a role. They only have to play the part well
@@LCx100 I don't see why someone would deep fake clothes, faces yes but clothes... That don't seem like a realistic evolution unless you talking about mock-ups (less money on actors and more profit) which that already exist
So here's how it works. Business decides that some technology is useful for them and saves costs, and then the rest of society is just shouldered with whatever consequences that technology brings with it.
Although quite often you can notice a deep fake still, as there are some tell tail signs, in just a few years it will be very difficult to pick the real from the fake apart. I would say that we could, pretty much, at this moment, have A.I create a story, write a script for it, create the sets and people for each scene using the looks of real actors and actresses, pose and animate them, create the CGI effects and create a score to go with it. In essence, create a complete movie from start to finish. The only thing at the moment is that some improvements are need to be made in some areas so that you can't tell the difference from the real and fake versions but i think we're rather close now. So the future seems it will become one where you don't even have to have actors, directors, CG artist, and so on to create a movie. You will simply get the A.I to generate all the parts needed, including the story and there you have a money maker from virtually nothing. You wouldn't put any work needed into it as none would be needed unless of course you wanted to create a specific style of movie but even then you would just tell it what you want. The rest would be created by A.I.
Deep Learning algorithms won't need until 2024 or 2025 to show perfect fakes. in many ways, the best deeps are already impossible to tell apart with our naked eye. Deep Learning algorithms are advancing in one month what it would've previously taken one year. Essentially, it's becoming impossible to keep up with the progress of these new age technologies.
No offense, but your prediction is extremely naive. AI has a huge problem that can be appreciated for what those initials stand for: Artificial Intelligence. If we invert the equation, we get: Natural Stupidity, or properly: Real Stupidity. Where am I going with this? Well, AI systems suffer from a flaw. They devolve from intricacy to route one tactics within a short period of time. For example, their initial offerings may seem authentic, realistic, convincing etc., but as they are given more autonomy, they dispense with details, and apply a path of the least resistance approach to a problem. For example, if you try using an AI art generator to hone in on a specific task or look, you'll quickly realize the system degenerates into offering offensive, inaccurate concoctions. To create a narrative, you need to have some way to parley verisimilitude. Only humans can do this. A machine, motivated by some let's say "special" logic, has no desire to produce something of value. It gravitates towards a goal. Think of it like a bolt of lightning. It simply follows the easiest path. Creating a narrative is an extremely complex undertaking. Making something satisfying is a whole other level of complexity. Then beyond this, there are many, many issues involved in creating CGI elements, models, textures, lighting, animation. I'm sorry, but there is no AI system that can take on those tasks autonomously, and it will be a very long time before such a system exists, if ever.
I guess things are just going to become less and less real as time goes on. This is the first time I've seen 'sci-fi' technology and been more concerned than amazed (even though it's still both.) I've known about deepfakes for awhile now, but...this video seemed to put it into a realistic enough context to make me go O_O.
The world in 10 years.... You use Neural link to log into your surrogate, and everything in life is sanitized, polished, and air brushed. Real life and authenticity are banished. We will live in a Black Mirror episode. or we will blow ourselves up. which I think is more likely.
From this moment on we will have to have honest remote viewers to sort through the muck of the world. Obviously if someone doesn’t like you they can make you look guilty.
The strangest thing about all this is; it turns out those 'subtle' face expressions you attribute to someone's unique personality, are not unique at all...Simply placing the digital face over someone else's face, seems to automatically capture all those small micro expressions....
There are benefits and drawbacks to everything. I appreciate you covering everything on DeepFake rather than making it seem like something that'd doom the whole world. Still a little alarming to see how advanced some of this technology is though. You can only wonder what the world will look like from here on out.
You are fucking naive. No other way to put it. Whatever "good" can ever come from any of this, it will not outweigh the life's that will be ruined, and deaths causes by this. Remember this children, if the risk is death (and it is, look at the Ukraine video) NOTHING is worth it no matter the "good".
One instance of deepfake I would love to see happen deals with the Disney movie Fantasia. When it was first released, the narrator was a guy named Deems Taylor. He had a very soothing and appealing voice. Later on, in other releases, another man's voice was dubbed over his. If you buy an old VHS copy, you can still hear Mr. Taylor's voice, but the video quality isn't good enough. In the newer releases, more narration scenes are added, but they had no audio of the voice from these segments. That's why another narrator was used. I would love to see deepfake tech used to update Fantasia with Deems Taylor's voice used everywhere. I would love this. Please.
The idea of a game that adapts to you and changes based on your actions just becomes more and more feasible. You won't even have to record a voice line for anything, all you need to do is set the entire game in the hands of the AI and all of a sudden every single player plays a different game.
I'm a computer scientist. The main problem is, this would require tremendous computing power. I just don't think it would be feasible with traditional computers, even in the future. There are limits to how good traditional computers can get which are imposed by the laws of physics, and we're already pushing right up against those limits. I think the only way to overcome that problem would be if we finally get quantum computers working properly, then we would have access to unimaginable levels of computing power, enough to make something like you described trivial. But there are quite a few issues that need to be solved with quantum computers before they will be useful, and although we are slowly making progress, some of those problems are very hard to solve, and are inherent to quantum physics. The other issue would be how do you ensure that what the AI shows the player is actually fun? Sure, it could probably build worlds, but without a human planning them out and deciding how they fit together, it would be very hard to teach an AI to build an actual game. A lot of the process of building games is creative, and creativity is something that AI really, really struggles with. That being said, DALL-E has proven that AI can be creative, and to be honest, I hope that I'm wrong about this. It's just, I think it won't happen for quite a long time, if at all. Although AI in games will get more and more complex, and maybe even begin to use methods like neural networks or deep learning.
It would be amazing to play a game like that. While there are games that do adapt and change, it is still limited and also partially scripted at present. I've yet to play it, but Echo comes to mind, which reacts to how you play it. That said, being able to hear characters say the name you chose for your character would be nice to hear. But we're likely a long way off these kind of games, like CynicallyDepressed says, our current technology isn't able to reach it. At least for the standard gaming PC or console.
With cancel culture rampant, it would be super easy to make a deep fake of some celebrity and have them say some outlandish things that will ruin their reputation and career, without that celebrity actually being involved in any way. And that's literally just on the lower scale of things that could go wrong. 😑
Its scary to think that the kind of future Cyberpunk 2077 predicts might not be that far away when I look at how AI imagery and deepfakes are progressing.
@@craid23 no you didn't. Not exactly AI imagery literally. But in an uneducated estimation, a very evolved version of combination of tech like this. Remember there's a whole side plot about Peralez couple.. what happens to Mr.Peralez. (being cryptic incase it's spoiler territory)
ปีที่แล้ว +6
I'm guessing deepfake detection software is only going to make deepfakes more accurate, as most probably those can be used to train deepfake software further.
Paramount Pictures needs to use deepfake technology to produce seasons 4 and 5 of Star Trek The Original Series. The five year mission needs to be completed and it would be great to see the original crew of the Enterprise in action again.
Can't believe this level of content clear.. by the time ii see this it has 328 views,75 subscribers,0 comments,12 likes in 3 days (thank goodness i see this video) Now let me add 1 count while sharing & trigger the youtube algorithms .. wishes you reach more soon bro 😃🌟✨🙌 (Ps:i am not a bot 😅) 13.09.2022 07:41 pm ist
Wow, audio deepfakes would be really useful for localizing games into less widespread languages. Maybe, they can even be used by fans to create some cool mixes of their favourite performers' voices.
thats what i was thinking, combine it with decent AI story generation like Ai dungeon and you could have infinite voiced story quests in Games, man the possibilities are endless..
end of this decade they will be completely indistinguishable. You have to depend on official sources that are vetted to make sure the content is supposed to be legitimate, and then of course run apps to detect if it is a deepfake and hope they can catch it if it is. Considering democracies globally are struggling with misinformation now, it's pretty terrifying to think how hard this is going to be for the average person to filter through. Nobody is going to know what is real anymore online. It's just going to be a huge cesspool of fake stuff everywhere.
I like how you completely neglect to mention what people are most worried about when it comes to neural network generated content and that is pornography, I made a post on Reddit a couple of days ago about this asking if OnlyFans models would be replaced with these networks instead since I believe that in the near future you'll be able to take any photo of any person online be it an adult or a child and generate pornographic scenes of that individual using these networks, even beastiality should be possible. No longer will we have to rely on what these people are willing to do themselves, we just produce it ourselves. Maybe you should make a video about this, something that people are both disgusted and scared of but will be a major market for these type of tools.
I knew about it. DeepFake started its journey in the pornography world. I felt disgusted and I don't want people to pursue it so I completely neglected it.
I've seen some of these deepfake videos using celebrity faces. Bet some indie game developer is going to make VR games of similar nature with this tech, lol
Everyone talks about how easy it will be to convince people of stuff that isnt real. Nobody talks about how hard it will be to convince people of what actually is real. This WILL cause a global crisis and possibly a total breakdown of society. My dear friends, humanity is royally fugged :D
I don't know if they'll ever be perfectly able to fool everybody. Your brain is incredibly good at picking out and flagging anything "uncanny" as a sort of defense mechanism. The subtle differences, slightly off lighting, unnatural speaking patterns, and jarring movements are clearly obvious to me, at least. Who knows, though. Maybe in another ten years or so it will be impossible to tell the difference.
….Yknow the human brain is surprisingly as flawed as it is efficient. All a computer has to do is find out the solution mathematically and apply it. It’s only a matter of time before a person cannot tell.
Using Deepfake to open up the world to films recorded in only 1 language is an amazing leap forward in allowing people to experience other cultures, by being able to hear the original actors speaking in a language which they can understand, with lip-synch matching the words, it will engage people who cannot keep up with reading subtitles (such as dyslexics) and people who cannot tolerate dubbed films with lips not matching words (such as autistics), as well as bringing the original actors world wide fame. Also being able to allow computer games to have an infinite ability to communicate with players with real-person like avatars lip movement matching & human voice takes the depth of immersion to a new level of virtual reality gaming! And don't forget, the computer games industry makes MANY times more than the entire film industry every year, and has done so for quite a long time now! As with any tool, it could be misused, but in the same way that a knife could be used to kill someone, it is also necessary to feed people. So, the tool, deepfake, should be allowed to progress and be embraced within the creative industry. It is those end user individuals who try to misuse it who should be punished for their choice to abuse a tool which can bring joy to others. And a key point here - as the technology for Deepfake gets better, so will the software which can be used to identify a video has been altered, & the use of "check its real and unaltered" software will become widely available for people to use, if they worry that some footage has been tampered with, in the same way that there is software which can spot when a photo has been photo-shopped.
There is one thing most people miss... As AIs advance, the gap between AI made videos and real recordings will decrease. once an AI can create a video so accurate that it cant be distinguished from real video, there will simply be no way to tell them apart. given how fast the tech has come, within 1-2 years it should be easily possible for simple video. more complicated for larger scenes, but within 5 years i doubt there will be any way to tell them apart provided you can access it. if there is money to be made then we will surely be able to.
@@jessiejanson1528 this is a great thing for making amazing new films. It will never be a security risk as the technology to spot a deepfake is always progressing alongside the deepfake itself - as one improves, so does the other.
@@TimLongson the reason you can tell fakes are fake, is because they arnt perfect. there will come a day where the fakes are perfect. At that time you wont be able to tell them apart. unless you want to suggest that in 1 million years we still wont be able to do this, then the question isnt 'is it possible' but 'when' and given how fast this tech is advancing, the when is 'vary soon'.
This could literally drive people insane. If it's not in person, you wouldn't be able to trust anything you see on the news, internet, anywhere. Puts the old World of the Worlds radio broadcast to shame.
@@Damesanglante Or if half (or all) the news we see is fabricated or altered in some way. Change a face on someone at a crime scene and before you know it someone's life is destroyed. Even if the truth comes out, damage done. Put's the Matrix to shame.
I'm very interested in Audio deep fakes. Now we can finally 'fix' movies and series where the producers couldn't be bothered to get the same voice actor again for a sequel or season.
Had a teacher once tell our class that human beings are too smart for their own good. We were discussing the first nuclear bomb and other technology that humans have created. It's scary that our own inventions will likely lead to our own destruction one day.
I remember when misrepresentation and falsifying ones identify was considered fraud... Even worse is that it wasn't that long ago that we were educating children to be aware of predators disguising their voice, now they can disguise their image... Advancing technology is one thing but this needs to be outlawed!
We need to keep ahead of this technology and encourage people to access the already existing tools that allow us to determine if a video is fake or not.
I hate this whole concept. Not just for the global political and moral ramifications of powerful new technology (ever heard of Black Powder? created to make mining and clearing of obstalces for infrastructure, and also for entertainment like fireworks...This tech is basically just "digital Black Powder..."). I'm also against this as automation of art, just like the new AI generated artworks currently gaining popularity online will only lead to a dystopia of "controlled art" by whomever holds the programming of these AI and also, if left for long enough - less people developing those skills and less human created artwork... Because why do a job when a machine can do it for you, right?? There is a point where automation becomes too much and unacceptable, and removing the humanity from artistic creation is past that point for me. Only way it will become acceptable is if AI is self aware and holds some kind of "True Artificial Life" status and real emotions and feelings as humans have. And that level of AI in itself should scare everyone. That could potentially be humanity engineering it's own demise. Natural selection will favoured the most robust organism, and the second we make True AI Lifeforms, at least as good as ourselves, that is truly just walking ourselves up to the executioners block.
“This should increase the variety of entertainment and content that we consume” - this is the second biggest problem with this and AI tech in general. The flood of stuff to consume will leave us, as an audience, numb and uninterested, while leaving us, as creators, incapable of making a living because there will be way too much stuff available (one could argue that is already a reality) for our work to get any sort of traction and generate relevant income. It will all become way too simple and too fast to achieve. This will implode on itself.
Just because we can doesn't mean we should. And ethics is definitely something to be considered. But it doesn't really matter anymore does it. The cats out of the bag. The technology exists and it can't be undone. If you really want to paint deep fakes in a good light then you should continue to educate and also push for the moral and ethical uses. Like a gun it only does harm when the user intends to.
We're in for a time when it's less important who said what, but it's becoming more important what's being said. Even deep faked content is human made. But I guess it's going to pose a treat, right. I need to make a correction. I made a critical typo in the last sentence. I meant to write the following: *But I guess it's going to pose a threat, right.*
I didn't even notice his voice was deepfake before I read the comments... This is very disturbing. Also the videos are getting near perfect too. I am concerned, maybe even scared. If you are not, you probably aren't paying attention enough.
Wow, this is a very potent technology and multitude of uses. But I'm still concerned about the moral issues it can cause, like crime and catfishing. In the meantime, excuse me while I'm making some Baka Mitai deepfake compilation video
Nice to know that all actors will be out of a job in the future and the AI will put together any actors it wants for its movies, naming them slightly different from the real actor.
People don't realize how dangerous this is. Anyone can take pictures of you and use that other AI to copy your voice and make you say whatever they want. This is chilling
Somehow I expected this long ago. It's why I don't like video surveillance.
i wonder if we will have a counter software to detect deepfakes
@@guimcast1 by the time we prove it's deep fake the WWIII will already have begun.
I forget the book but people would have slanderous content produced about them in volume to make any real allegations less believable.
@@guimcast1 in 10-20 years they will probably be completely indistinguishable from a real video. Instead of developing software for identifying fake videos, we'll probably have to go the opposite route and validate real videos through a token system or something like that
The fact that the the audio used in this video has also generated with the help of AI which makes it even more concerning.
I just realized it when he pronounced 'AI' wrong. Didn't even notice at first. Very impressive.
I also noted that the way he said A.i is a bit weird, but i had no reason to believe the voice over was fake. I simply scoffed it up to the British accent making A.i sound weird. I'm a bit shocked.
what software did he use?
@@coomer_artist2648 I think speechify
I'm excited and terrified because this shit is coming fast a furious now. Machine learning and AI-generated everything. Just like they said it would, just like we all knew it would, but now that it's here and the Boston Dynamic robots can run around like people, and the humans can't tell if they're talking to a person or a computer and even my PHONE - an iPHONE 7 is showing much greater levels of awareness than it did a few years ago (same phone, updated OS).
AI-generated images are fucking insane now. A few year's ago, they were not even abstract art, they were just nonsense vomit. Literally two years ago. now, they are photorealistic. Within next ten years, we'lll have full length movies that are 100% generated by AI.
But real intelligence all around, genreal intelligence will still only exist in a few super computers. Those super computers are borderline sentient, though. That "jedi" joke by the Google super computer AI is funny, witty, clever, and clearly a level of understanding that exceeds that of my five year old.
If they're openly telling us about it, just imagine what's going on.
Thas rite!
Once you go down the rabbit hole there is no coming back
@@nw7654 the problem with the rabbit hole is finding truth amongst all the lies
@@Sin.Atonement That is pretty much everyday life unfortunately.
@@nw7654 yup, black magic all around us
Man, even your voiceover had me convinced it was being delivered by a real person! I didn't catch on that it was AI until you had to pronounce "debatable" haha
Yeah😆
This actually scared me. I thought the voice was real. I heard the breathing and was totally convinced. Took me until 1:18 when he didn't pronounce the I in Deepbrain AI. Although I should've noticed when he immediately mispronounced Simon Cowell.
Took me till simon cowel.
@@zohzie Yeah, haha, fair enough
@@exonet51 Exonet's reply is from an AI as well. God damn
We already live in a world where people lack the ability to distinguish fact from fiction. We're not ready for this.
That’s what I think too
Thats why AI will feed us "truth" it deems necesery, total monopoly over information, you cant use your "critical thinking" if objective facts are indestinguishable from deepfakes, Metal Gear Solid 2 predicted future 20 years ago
We’ll never be ready. Too many evil people in the world!
Prophet muhammad already told this 1400 years ago..were on the last age..
If you know what you're looking/listening for, you can still spot AI generated content. Examples:
1:16 Deepbrain Ayy
1:58 De-bat-able
4:29 High perr-ealistic
5:15 Sensity is creator~ platform
6:04 A similar initiative _bylays_
Yeah, AI can't pronounce simple words, it's far from destroying our world.
@@exonet51 At least a year or two away
Simon Coal
As english is not my first language I would have never spotted that. Rather scary.
@@exonet51 "far" is relative. These pronunciation errors will be fixed in a few years and the tell tale signs will become ever more subtle to the point where the vast majority of people won't pick them up!
We already live in a world where half the people can’t agree on objective reality. With this technology, within a few years, there will be no consensus whatsoever on reality anymore. I shudder to think of the ramifications.
Just read books, ignore tech. Easy life.
No one has ever known reality. But when people don't know who's "real", we're fucked... And good luck convincing the "not" real ones. Fortunately, they're not in the "real" world yet...
Okay, I was going to sleep and now I've generated an existential cricis. YES!
@@feboptopt but you are not doing it yourself since you are on TH-cam.
@@willy4170 you can do both.
@@trulymental7651 but if you do both you are not “ignoring tech” so get off youtube and go read some books, those who say those things, shouldn’t be on any social media to begin with, but just “ignoring technology”
It's not so much that people will start believing anything; it's that they'll stop believing anything 🤔
Well they should… we have been lied to by doctors and politicians way before
In a world of lies, trust is becoming the new currency.
It's the only thing which matters, actually. This is why Huge Companies are now dominating Ecommerce. The public is paranoid, using anybody they have never heard of online. So Google, eBay, Amazon, Walmart dominate sales. Building any of these is not hard, but without the Brand it's pointless.
In a world of lies, trust is becoming impossible. As a currency, trust works as well as unicorn hair.
I dread the day where someone has evidence of someone doing something and that person doesn't get in trouble because they can't "prove it wasn't a deepfake".
by that point no one will take video seriously anymore anyway .... because of this ...... after all it`s already a century old technology and was bound to be compromised eventually
@@noiselabproject9659 metadata will probably be used to some extent to separate the one from the other
Or the other way around, innocent people getting locked up because of video evidence!
@@EyelessEntity The problem with that is that even metadata can be tampered with. Essentially we will be getting to the point, as was pointed out, that only another AI will be able to distinguish between real and created by another AI, which ironically would provide the algorithms for the fakes to be made even more realistic. This is a snowball that has just begun to roll downhill.
@@Dilligff You've just described an arms race. Everyone has to 1up the other guy.
Taking On AI | Mission: Impossible - Dead Reckoning Part One: th-cam.com/video/HF8BcUBUBb4/w-d-xo.html
There are a few moments where the intonation either drops or raises weirdly, but you could easily write this off as a poorly edited cut in the audio. This was terrifyingly convincing
I couldn't even notice.
… Yeah. This is both cool yet makes me INCREDIBLY concerned about news reporting, propaganda, and public opinion influencers
ALSO LETS NOT FORGET HOW 70% OF TH-cam COMMENTS ARE FAKE
Yeah but give it 3,5,10 years. Where will it be then?
When you watch a TH-cam video, you are ready for those short "cuts" and "dents" in recording, as creators don't need to use the smoothest of the most refined film montage. That's the main factor allowing most AI talk failures to pass under the radar.
Especially scary since the majority of people aren’t as aware of the nuances of what makes something detectable as real vs fake. Most people can’t even tell when something is OBVIOUSLY cgi or photoshopped. Most people don’t pay attention or see reality, they see what they predict they’re seeing.
True.
Its the difference between active and passive observation.
3:44 🤣
And what's even scarier is that some people WANT to believe false information about people they don't like and won't be convinced it's fake even when presented with overwhelming evidence...
Yeah, I dread the day where someone has evidence of someone doing something and that person doesn't get in trouble because they can't "prove it wasn't a deepfake".
Been looking for a channel like this for a while... non-sensationalist and succinct info on AI updates, and potential scenarios, without jargon and without cheap thrills. Keep it up.
Thank you for your support
This channel still has clickbait thumbnail though, not that there's big problem with that, but still. Especially this video where thumbnail image never appeared in the whole video itself (1 frame lasts like... 1/30th second does not count, of course)
@@3enjamin yeah, sometimes it's kind of annoying but I got what I wanted from the video.
@@3enjamin That's true, we should remind the creator of that in a new video!
@Patrick Silva There's been good and bad things going on simultaneously since the dawn of time. It's an eternal struggle. We can only support the good guys, hope for the best and be ready for the worst.
What is terrifying me the most is that the database of almost everyone using social media is already made to be used.. any of our close friends or family is a target if the tech easily reaches hands..
Hence why I came off all social media 7 years ago…. I saw this as threat then and it’s become very apparent now that this technology opens a bigger worldwide problem.
This video was made by AI.
The problem is that as technology improves they might find a way to make it undetectable.
Sure thing. If the deepfake reaches the point of flawlessness there is no way to detect it. Even with tools.
Nah I disagree, detection tools that work like anti-virus tools will most likely improve and develop faster than deep fake technology. Look at what happened to viruses on PC’s. Viruses were rampant in the early 2000’s. But as big money was being pumped into research for anti-virus technology over the years, viruses are not that common anymore. With Bidefender along with taking other easy security measures your pretty safe, and theres still better anti-virus software than that that you can buy.
That is why, developing critical thinking is very important. Verification before jumping to any conclusion is an essential skill.
This is why we do not trust media.
Yeah but some media are thinking to use it for the right cause.
When I was young I used to think robots would take over all of the laborious work leaving humans to explore the arts and science. Now that I’m older I I’ve come to know that the robots are here to entertain us and I still have to mow my own lawn.
Brings a whole new light to the old saying that taking a photo of someone was stealing their soul. Now, in the wrong hands, it could easily be just that.
How true . In which culture or teaching or country this saying belongs too?
Islam discouraged people to take photos of living beings even if it is a selfie . And any drawing or picture of any creature with eyes are considered a living object . Its also discouraged. As the time goes through history , now I am understanding the wisdom of it . Alhamdulillah
It never fails to amaze me the lengths we will go to to make ourselves obsolete...
Not we, them...
There is a difference
I believe you are right on the button Laura..🎯. Somehow I have this odd feeling that the universe is Teeming with life, & at the same time its a graveyard littered with the evidence of species that innovated themselves right out of existance.
Looking forward to seeing movies with old actors again on the screen...
This is the type of stuff that was considered science fiction not that long ago. There's still an uncanny valley but it's getting scary close. I'm reminded of the Michael Crichton novel Rising Sun which had video being altered to frame someone else for a murder caught on video. It's a cool tool and when used creatively can do wonders for film and entertainment. But in the wrong hands, this could cause some serious trouble.
That's what more people should be focusing on...what this will be used for in the wrong hands. Or, perhaps, it was always intended for nefarious purposes and things like movies are just the cover.
There's a lot of stuff we've created that is actually amazing. Look at the internet as well as social media. Great tools, yet often greatly abused. Deepfakes are just another thing, amazing technology, but what are the consequences of using it.
And now people are wishing they had paid attention to that sci-fi instead of calling the readers dorks.
It's already convincing. Look at Luke in The Book of Boba Fett.
Indian Boy British Accent 😎 smooth asf...Top notch content bro...I wonder which AI Voiceover is that!! It took me 6 videos to realise,its AI 😂
It's Wellsaid😄
This is a very concerning issue. I think that we should have some type of international law that protects likeness and voice prints.
Also, if a voice print is used, the original voice actor should be compensated. The could be a very rewarding process for many if it is done ethically.
This is one of those things that just because we can do it we shouldn't. Way more bad things come out of this than good.
My thought extactly like that too, scary o.o
Not true. This tech will be instrumental in creating digital realities. The downsides are literally trivial by comparison, especially as they are not actually that hard to deal with. Why everyone acts like a fake speech by a president will start WW3, I don't know.
The future is now old man!
Why? Did you always trust what you saw on TV before?
@rasmasyean no, but a lot of people are stupid anymore and do believe everything they see and hear. Kinda like you. Someone who assumes and jumps all in thinking they know everything.
Now we can safely conclude that
Tom cruise is the king of deepfake.
We will have long be gone and he will be still making movies.
They could at least make him appear
to be taller than a garden gnome.🤡
Plot twist: Tom Cruise died three years ago, and everything we've seen of him since has been fake.
@@marendur well ... he won't be making them, but he'll be in them.
One way to counter this would be. You would be forced to have your own website to post an "Official" release. New outlets would need to embed a link from your source, any other would be considered unreliable.
Yeah... We just had nations lie about how a lab made virus was released...
Nations and corporations won't adhere to "best use" ethics.
They'll push us all to war in order to further advance this tech, along with AI, brain implants, human augmentations, and social credit scores.
This is not heading in a good direction for humanity... Not one bit!
@@angryherbalgerbil I clicked "Read more" but then clicked "Show Less" less is better. Good luck off your meds
I thought about deep fakes for a couple of years now. With the state of the world as it is the nefarious possibilities make me nervous.
What if this is primarily FOR nefarious purposes and movies n such are just a nice way of getting people accustomed to the idea of it...
@@katie7748 I think you're on to something. I can't imagine why people would even create deceptive tech like this unless they planned to do terrible things with it right from the beginning.
We were so preoccupied with whether we could we didnt stop to think if we should.
………wow…..how original.
@@mw9688 nothing new under sun, sweet cheeks.
That's right. It's not like our conscience doesn't warn us or that the alarm isn't sounded along the way. Jurassic Park quote, right?
@nostromo79 yeah, Jeff goldblum's character talking about bringing back the dinosaurs.
With the rise of deepfake and AI generated images, actors of the future won't have to look a particular way to fill a role. They only have to play the part well
Yeah, that's the advantage of using deepfake
Makeup artist may go jobless with this tech
Hell they may not require any cellebrity to remain present on stage for this
@@LCx100 I don't think so, you can't deep fake makeup on clothes and clothes need makeup believe it or not
@@deadplthebadass21 give it some years that fault will vanish too
@@LCx100 I don't see why someone would deep fake clothes, faces yes but clothes... That don't seem like a realistic evolution unless you talking about mock-ups (less money on actors and more profit) which that already exist
So here's how it works. Business decides that some technology is useful for them and saves costs, and then the rest of society is just shouldered with whatever consequences that technology brings with it.
This is the last thing we need in times like this.
1984
That is why you believe non of what you hear, and half of what you see.
I first learned of audio deepfakes from the Mark Hamill one, honestly I think audio deepfakes are even more technically impressive than visual ones.
There's a saying:
When you have no deeper sense of value, you start taking everything at face value.
Although quite often you can notice a deep fake still, as there are some tell tail signs, in just a few years it will be very difficult to pick the real from the fake apart.
I would say that we could, pretty much, at this moment, have A.I create a story, write a script for it, create the sets and people for each scene using the looks of real actors and actresses, pose and animate them, create the CGI effects and create a score to go with it.
In essence, create a complete movie from start to finish.
The only thing at the moment is that some improvements are need to be made in some areas so that you can't tell the difference from the real and fake versions but i think we're rather close now.
So the future seems it will become one where you don't even have to have actors, directors, CG artist, and so on to create a movie.
You will simply get the A.I to generate all the parts needed, including the story and there you have a money maker from virtually nothing.
You wouldn't put any work needed into it as none would be needed unless of course you wanted to create a specific style of movie but even then you would just tell it what you want.
The rest would be created by A.I.
Yeah, That will be a great application of AI and Deepfake.
Deep Learning algorithms won't need until 2024 or 2025 to show perfect fakes. in many ways, the best deeps are already impossible to tell apart with our naked eye. Deep Learning algorithms are advancing in one month what it would've previously taken one year. Essentially, it's becoming impossible to keep up with the progress of these new age technologies.
It won’t make you any money when everyone can do it. :D No skill will be involved if an AI can do a full movie with a click of a button.
No offense, but your prediction is extremely naive. AI has a huge problem that can be appreciated for what those initials stand for: Artificial Intelligence. If we invert the equation, we get: Natural Stupidity, or properly: Real Stupidity. Where am I going with this?
Well, AI systems suffer from a flaw. They devolve from intricacy to route one tactics within a short period of time. For example, their initial offerings may seem authentic, realistic, convincing etc., but as they are given more autonomy, they dispense with details, and apply a path of the least resistance approach to a problem. For example, if you try using an AI art generator to hone in on a specific task or look, you'll quickly realize the system degenerates into offering offensive, inaccurate concoctions.
To create a narrative, you need to have some way to parley verisimilitude. Only humans can do this. A machine, motivated by some let's say "special" logic, has no desire to produce something of value. It gravitates towards a goal. Think of it like a bolt of lightning. It simply follows the easiest path.
Creating a narrative is an extremely complex undertaking. Making something satisfying is a whole other level of complexity. Then beyond this, there are many, many issues involved in creating CGI elements, models, textures, lighting, animation. I'm sorry, but there is no AI system that can take on those tasks autonomously, and it will be a very long time before such a system exists, if ever.
@@screenname1 very long time, so like 5 years tops
I guess things are just going to become less and less real as time goes on. This is the first time I've seen 'sci-fi' technology and been more concerned than amazed (even though it's still both.)
I've known about deepfakes for awhile now, but...this video seemed to put it into a realistic enough context to make me go O_O.
The world in 10 years....
You use Neural link to log into your surrogate, and everything in life is sanitized, polished, and air brushed. Real life and authenticity are banished.
We will live in a Black Mirror episode.
or we will blow ourselves up. which I think is more likely.
We are going to enter the age of real deception where humans are not valued and truth is barely seen
what did you use to produce the voice on this video? It's almost perfect!!
I used a website called Wellsaid
@@exonet51 Well thank you!!
thats a very convincing audio bot v/o, i thought it was a real person at first
Yeah, it also fooled at first😆
@@exonet51 oops, this reply needs work
From this moment on we will have to have honest remote viewers to sort through the muck of the world. Obviously if someone doesn’t like you they can make you look guilty.
The strangest thing about all this is; it turns out those 'subtle' face expressions you attribute to someone's unique personality, are not unique at all...Simply placing the digital face over someone else's face, seems to automatically capture all those small micro expressions....
They are unique but you actually need to know someone well to notice
There are benefits and drawbacks to everything. I appreciate you covering everything on DeepFake rather than making it seem like something that'd doom the whole world. Still a little alarming to see how advanced some of this technology is though. You can only wonder what the world will look like from here on out.
I think AI will be able to spot faked videos even better than produce it.
The risks seem so much higher than the benefits in this case though
You are fucking naive. No other way to put it. Whatever "good" can ever come from any of this, it will not outweigh the life's that will be ruined, and deaths causes by this. Remember this children, if the risk is death (and it is, look at the Ukraine video) NOTHING is worth it no matter the "good".
The drawbacks (being cataclysmic in scale) outweigh the benefits by a landslide though. This is a pandoras box.
First thing I read in the thumbnail:
"THIS IS FAKE"
First thing I thought:
"Oh but they look natural!"
*reads title*
"Ohhh!"
One instance of deepfake I would love to see happen deals with the Disney movie Fantasia. When it was first released, the narrator was a guy named Deems Taylor. He had a very soothing and appealing voice. Later on, in other releases, another man's voice was dubbed over his. If you buy an old VHS copy, you can still hear Mr. Taylor's voice, but the video quality isn't good enough. In the newer releases, more narration scenes are added, but they had no audio of the voice from these segments. That's why another narrator was used. I would love to see deepfake tech used to update Fantasia with Deems Taylor's voice used everywhere. I would love this. Please.
We are reaching an age where it's hard to tell what's real or fake, who to believe or not to believe
@happy little trees 239 can’t trust that either
Question everything.
Already there and always have been.
The idea of a game that adapts to you and changes based on your actions just becomes more and more feasible. You won't even have to record a voice line for anything, all you need to do is set the entire game in the hands of the AI and all of a sudden every single player plays a different game.
I'm a computer scientist. The main problem is, this would require tremendous computing power. I just don't think it would be feasible with traditional computers, even in the future. There are limits to how good traditional computers can get which are imposed by the laws of physics, and we're already pushing right up against those limits. I think the only way to overcome that problem would be if we finally get quantum computers working properly, then we would have access to unimaginable levels of computing power, enough to make something like you described trivial. But there are quite a few issues that need to be solved with quantum computers before they will be useful, and although we are slowly making progress, some of those problems are very hard to solve, and are inherent to quantum physics. The other issue would be how do you ensure that what the AI shows the player is actually fun? Sure, it could probably build worlds, but without a human planning them out and deciding how they fit together, it would be very hard to teach an AI to build an actual game. A lot of the process of building games is creative, and creativity is something that AI really, really struggles with. That being said, DALL-E has proven that AI can be creative, and to be honest, I hope that I'm wrong about this. It's just, I think it won't happen for quite a long time, if at all. Although AI in games will get more and more complex, and maybe even begin to use methods like neural networks or deep learning.
Imagine a game where upload your voice and images of yourself so it can be replicated on screen. That’s just immersive.
@@alphaohenriquez That actually could be done, it would still take a lot of computing power but I think it would be possible even today.
It would be amazing to play a game like that. While there are games that do adapt and change, it is still limited and also partially scripted at present. I've yet to play it, but Echo comes to mind, which reacts to how you play it.
That said, being able to hear characters say the name you chose for your character would be nice to hear. But we're likely a long way off these kind of games, like CynicallyDepressed says, our current technology isn't able to reach it. At least for the standard gaming PC or console.
@@CynicallyDepressedx true
people think anyone can Deepfake, no!
With cancel culture rampant, it would be super easy to make a deep fake of some celebrity and have them say some outlandish things that will ruin their reputation and career, without that celebrity actually being involved in any way. And that's literally just on the lower scale of things that could go wrong. 😑
Its scary to think that the kind of future Cyberpunk 2077 predicts might not be that far away when I look at how AI imagery and deepfakes are progressing.
Huh? I can't remember Cyberpunk 2077 touching on AI imagery and deepfakes... or am I missing something?
@@craid23 no you didn't. Not exactly AI imagery literally. But in an uneducated estimation, a very evolved version of combination of tech like this. Remember there's a whole side plot about Peralez couple.. what happens to Mr.Peralez. (being cryptic incase it's spoiler territory)
I'm guessing deepfake detection software is only going to make deepfakes more accurate, as most probably those can be used to train deepfake software further.
Truly, you will not be able to believe anything that you see or hear on any kind of media outlet.
When you realise that the voiceover is not real ! LOL
Paramount Pictures needs to use deepfake technology to produce seasons 4 and 5 of Star Trek The Original Series. The five year mission needs to be completed and it would be great to see the original crew of the Enterprise in action again.
Yeah, that's where DeepFake really makes sense
Fans have finished the Original Series, check out Star Trek Continues.
I wonder when the gov/CIA/NASA version of CGI will go into general release?
Can't believe this level of content clear.. by the time ii see this it has 328 views,75 subscribers,0 comments,12 likes in 3 days (thank goodness i see this video)
Now let me add 1 count while sharing & trigger the youtube algorithms ..
wishes you reach more soon bro 😃🌟✨🙌
(Ps:i am not a bot 😅)
13.09.2022 07:41 pm ist
Thank you for your support😊 I understood you are not a bot but there are some bots in the comment section, I need to clear it
This is just plain dangerous. All is always fun and games until someone with your build, face and voice is caught murdering in a CCTV camera 😂
It could even be worse. This shit could be used to start wars if done right.
Great video! Thanks for putting this together.
Wow, audio deepfakes would be really useful for localizing games into less widespread languages. Maybe, they can even be used by fans to create some cool mixes of their favourite performers' voices.
hmm, isnt that illegal? idk
thats what i was thinking, combine it with decent AI story generation like Ai dungeon and you could have infinite voiced story quests in Games, man the possibilities are endless..
@@captainalex157 try novelai, much better than aidungeon.
Considering AI generated voice has come THIS far, I have no doubts that deep fakes will reach beyond detectable levels soon.
end of this decade they will be completely indistinguishable. You have to depend on official sources that are vetted to make sure the content is supposed to be legitimate, and then of course run apps to detect if it is a deepfake and hope they can catch it if it is. Considering democracies globally are struggling with misinformation now, it's pretty terrifying to think how hard this is going to be for the average person to filter through. Nobody is going to know what is real anymore online. It's just going to be a huge cesspool of fake stuff everywhere.
I on my channel use AI voice xd
You can still easily spot live deep-fakes by asking them to look left or right, but for pre-rendered video it is a lot harder.
I like how you completely neglect to mention what people are most worried about when it comes to neural network generated content and that is pornography, I made a post on Reddit a couple of days ago about this asking if OnlyFans models would be replaced with these networks instead since I believe that in the near future you'll be able to take any photo of any person online be it an adult or a child and generate pornographic scenes of that individual using these networks, even beastiality should be possible.
No longer will we have to rely on what these people are willing to do themselves, we just produce it ourselves. Maybe you should make a video about this, something that people are both disgusted and scared of but will be a major market for these type of tools.
I knew about it. DeepFake started its journey in the pornography world. I felt disgusted and I don't want people to pursue it so I completely neglected it.
@@exonet51 ew... Suddenly the term deep fake has a whole new meaning
I've seen some of these deepfake videos using celebrity faces. Bet some indie game developer is going to make VR games of similar nature with this tech, lol
For good or ill, the pornography industry has always been a major driving force in technology, from photography to virtual reality and now AI.
@@exonet51 AI can feel disgust ?
If I can ask, what software did you use for this voice? It sounds good!
I used Wellsaid
@@exonet51 thank you!
They used to say "believe half you see and nothing you hear" I guess we reached the "Believe nothing at all" point
or theres should be a regulation that if a video or picture is deepfaked, there should be a watermark or something to show that it is a deepfake
Everyone talks about how easy it will be to convince people of stuff that isnt real.
Nobody talks about how hard it will be to convince people of what actually is real.
This WILL cause a global crisis and possibly a total breakdown of society.
My dear friends, humanity is royally fugged :D
F....I was just playing cyberpunk. And the scary thing is... this gives the same vibes.
I don't know if they'll ever be perfectly able to fool everybody. Your brain is incredibly good at picking out and flagging anything "uncanny" as a sort of defense mechanism. The subtle differences, slightly off lighting, unnatural speaking patterns, and jarring movements are clearly obvious to me, at least. Who knows, though. Maybe in another ten years or so it will be impossible to tell the difference.
….Yknow the human brain is surprisingly as flawed as it is efficient. All a computer has to do is find out the solution mathematically and apply it. It’s only a matter of time before a person cannot tell.
@U G too bad… it’s been created and is here to say and improve in time.
@U G if you stop the company that doesn’t stop programmers in the outside that already have the tech and means to develop it further.
@U G well…only if you see it as defeat.
I don’t.
Using Deepfake to open up the world to films recorded in only 1 language is an amazing leap forward in allowing people to experience other cultures, by being able to hear the original actors speaking in a language which they can understand, with lip-synch matching the words, it will engage people who cannot keep up with reading subtitles (such as dyslexics) and people who cannot tolerate dubbed films with lips not matching words (such as autistics), as well as bringing the original actors world wide fame. Also being able to allow computer games to have an infinite ability to communicate with players with real-person like avatars lip movement matching & human voice takes the depth of immersion to a new level of virtual reality gaming! And don't forget, the computer games industry makes MANY times more than the entire film industry every year, and has done so for quite a long time now!
As with any tool, it could be misused, but in the same way that a knife could be used to kill someone, it is also necessary to feed people. So, the tool, deepfake, should be allowed to progress and be embraced within the creative industry. It is those end user individuals who try to misuse it who should be punished for their choice to abuse a tool which can bring joy to others.
And a key point here - as the technology for Deepfake gets better, so will the software which can be used to identify a video has been altered, & the use of "check its real and unaltered" software will become widely available for people to use, if they worry that some footage has been tampered with, in the same way that there is software which can spot when a photo has been photo-shopped.
There is one thing most people miss... As AIs advance, the gap between AI made videos and real recordings will decrease. once an AI can create a video so accurate that it cant be distinguished from real video, there will simply be no way to tell them apart. given how fast the tech has come, within 1-2 years it should be easily possible for simple video. more complicated for larger scenes, but within 5 years i doubt there will be any way to tell them apart provided you can access it. if there is money to be made then we will surely be able to.
@@jessiejanson1528 this is a great thing for making amazing new films. It will never be a security risk as the technology to spot a deepfake is always progressing alongside the deepfake itself - as one improves, so does the other.
@@TimLongson the reason you can tell fakes are fake, is because they arnt perfect. there will come a day where the fakes are perfect. At that time you wont be able to tell them apart. unless you want to suggest that in 1 million years we still wont be able to do this, then the question isnt 'is it possible' but 'when' and given how fast this tech is advancing, the when is 'vary soon'.
@@jessiejanson1528 HUMANS may not be able to spot fakes, but AI knows what to look for and can spot fakes every time - you can't kid a kidder.
@@TimLongsonso you are prioritizing entertainment over sanity?
This could literally drive people insane. If it's not in person, you wouldn't be able to trust anything you see on the news, internet, anywhere.
Puts the old World of the Worlds radio broadcast to shame.
@@Damesanglante Or if half (or all) the news we see is fabricated or altered in some way. Change a face on someone at a crime scene and before you know it someone's life is destroyed. Even if the truth comes out, damage done. Put's the Matrix to shame.
I'm very interested in Audio deep fakes. Now we can finally 'fix' movies and series where the producers couldn't be bothered to get the same voice actor again for a sequel or season.
Had a teacher once tell our class that human beings are too smart for their own good. We were discussing the first nuclear bomb and other technology that humans have created. It's scary that our own inventions will likely lead to our own destruction one day.
Or not.
@@cloudstorage9928 I'm not sure about that, Boston University created more deadly strain of Covid recently.
I remember when misrepresentation and falsifying ones identify was considered fraud... Even worse is that it wasn't that long ago that we were educating children to be aware of predators disguising their voice, now they can disguise their image... Advancing technology is one thing but this needs to be outlawed!
The news is already fake my dude
Exactly
Operation Mockingbird
Very informative video. Thank you.
Imagine the danger of stopping wars...The horror.
In the world of lies and madness, everything is turned into a weapon rather than a tool.
Nah, AI can't fool me if I don't watch it! :D
That's another level😆
We need to keep ahead of this technology and encourage people to access the already existing tools that allow us to determine if a video is fake or not.
This is what the BBC series "The Capture" is all about. Absolutely terrifying.
I hate this whole concept. Not just for the global political and moral ramifications of powerful new technology (ever heard of Black Powder? created to make mining and clearing of obstalces for infrastructure, and also for entertainment like fireworks...This tech is basically just "digital Black Powder..."). I'm also against this as automation of art, just like the new AI generated artworks currently gaining popularity online will only lead to a dystopia of "controlled art" by whomever holds the programming of these AI and also, if left for long enough - less people developing those skills and less human created artwork... Because why do a job when a machine can do it for you, right?? There is a point where automation becomes too much and unacceptable, and removing the humanity from artistic creation is past that point for me. Only way it will become acceptable is if AI is self aware and holds some kind of "True Artificial Life" status and real emotions and feelings as humans have. And that level of AI in itself should scare everyone. That could potentially be humanity engineering it's own demise. Natural selection will favoured the most robust organism, and the second we make True AI Lifeforms, at least as good as ourselves, that is truly just walking ourselves up to the executioners block.
To be fair, many animals that we could easily make go extinct still exist.
But yeah I get what you say with AI Art. I hate that it exists. So much.
Great video ! We are in deep sh --- !! 😷
Yeah😆
“This should increase the variety of entertainment and content that we consume” - this is the second biggest problem with this and AI tech in general. The flood of stuff to consume will leave us, as an audience, numb and uninterested, while leaving us, as creators, incapable of making a living because there will be way too much stuff available (one could argue that is already a reality) for our work to get any sort of traction and generate relevant income. It will all become way too simple and too fast to achieve. This will implode on itself.
Protecting your identity and information just got a lot more serious
The adult videos are pretty good.
“If we don’t consider ethics, it’s quite doable and simple”
THANK YOU SO MUCH THIS WAS EXTREMELY HELPFUL :D
In 20yrs..we won't know What's Real
Just because we can doesn't mean we should. And ethics is definitely something to be considered. But it doesn't really matter anymore does it. The cats out of the bag. The technology exists and it can't be undone. If you really want to paint deep fakes in a good light then you should continue to educate and also push for the moral and ethical uses. Like a gun it only does harm when the user intends to.
Yeah, you made a valid point
We are all doomed.
This is good for Hollywood but Hollywood would probably never use it
We needed this on the new Matrix
This video was so life changing, keep up the good work
Hello, is this a bot?
@@exonet51 genuine
@@exonet51 hello i am bot. Are you a bottom or are you a bot?
We're in for a time when it's less important who said what, but it's becoming more important what's being said. Even deep faked content is human made. But I guess it's going to pose a treat, right.
I need to make a correction. I made a critical typo in the last sentence. I meant to write the following:
*But I guess it's going to pose a threat, right.*
@Jake R you are right, English is not my first language. Thanks to you I re-read my comment and found a bad typo. Thank you.
Reminds me of what Edgar Allan Poe said :
"Believe nothing you hear, and only one half that you see."
I didn't even notice his voice was deepfake before I read the comments... This is very disturbing. Also the videos are getting near perfect too. I am concerned, maybe even scared. If you are not, you probably aren't paying attention enough.
Wow, this is a very potent technology and multitude of uses. But I'm still concerned about the moral issues it can cause, like crime and catfishing.
In the meantime, excuse me while I'm making some Baka Mitai deepfake compilation video
Nice to know that all actors will be out of a job in the future and the AI will put together any actors it wants for its movies, naming them slightly different from the real actor.
I’ve also considered that someday you can watch a movie and choose which actors you want to star in it.
This very balanced video not all deep fakes are bad. It could be art. Good video mate.
Yeah, You are right, If used correctly, deepfake is a great gift, Thank you mate
@@exonet51 your welcome
A sprinkle of clickbait also do wonders
Technology is a tool that can be used for endless good, and bottomless evil.
and very appropriately, using AI generated audio.
we're close to a point where you cannot tell whats real or not anymore. very scary
Or if you're stupid and believe everyone.
It should most likely be a felonly offence to deep fake another person without their written consent.
It should be a felony to be a wage slave
The deepfakes are so real
Yeah, In the future, they will be more realistic.
This human world is getting more and more dangerous everyday in every way. We are doing to ourselves. No surprise.