Name: Brainstorm AI London 2024: The Future Of Synthetic Media
Uploaded: 2024-04-15
Duration: 18 min 27 s
Description: Victor Riparbelli, Co-founder and CEO, Synthesia Mati Staniszewski, Co-founder and CEO, ElevenLabs Moderator: Jeremy Kahn, FORTUNE

 Victor Riparbelli, Co-founder and CEO, Synthesia Mati Staniszewski, Co-founder and CEO, ElevenLabs Moderator: Jeremy Kahn, FORTUNE

Brainstorm AI London 2024: The Future Of Synthetic Media

Photography by Joe Maher Fortune Brainstorm AI

Conferences

Hi, welcome back. I hope you had a good break. Um I'm very excited to be joined with uh some of the, the CEO S of two of the hottest sort of start ups in the gene of A I space at the moment. Um And as you heard from Ellie, uh both working in synthetic media, which is maybe using tech to generate something, but maybe also using, you know, existing video, still images or voice to generate media and content to show you how this works. First, we've got two very exciting demos. First, we're going to go to Victor who's going to show you a little bit about how synthesia product works. Victor. Go ahead. Thank you so much. Yeah, I think we should give a visual in just a second here, but maybe before we jump into that. Uh So these are on a mission to make video easy for everyone and that's, that's something a lot of people have attempted before us. But we take kind of a new approach to this, right? We don't think of this as building smaller and more affordable cameras. We don't think of this as like slightly better editing apps that run in your phone. We're just building technology that eventually is going to be able to replace the entire physical production process of using studios, cameras, actors, microphones and all those stuff. We want to take that entire workflow and make it into something you can do entirely from behind your desk. Um Today, we're a fast platform and we help predominantly enterprise create more videos to communicate better with their stakeholders, which could be employees, could be customers. And I think we'll get a visual up here now of how it actually works and how easy it actually is to do. So this is the platform and in some ways it's sort of model a little bit on powerpoint. That's how easy we want to be able to, to do. And what we're showcasing here is our A I video assistant which uses LL MS to essentially help you get to a draft of your video in, in just a few moments. So what you're seeing here is someone putting in uh the URL to an article, of course, a fortune article because we're here today giving the system a little bit more context of what it is that we're trying to do with this video. And in just a second, the system will then go in, it will pass the content and it will give us a draft on editable copy that we can then kind of tune to what exactly we want it to be like now in the context of an enterprise, this may not be uh an article like this, this would probably be a knowledge based article. It could be a case study that you want to share with your customers or anything else really, you want to turn into a video as you can see here. Now, um we begin to actually write the script and it is of course using all the data that it's kind of taking from this URL. The idea here is not that this is a final video that you can just publish, but it is to take you 70 80% of the way. Instead of having what we internally call the blank screen of death, we give you something to actually begin kind of working from and we don't just write the script, we actually also give you the visuals. So as you'll see here, there's some text on the screen, there'll be some bullet points, there'll be a bunch of things that you can sort of edit yourself. Of course, everything can be edited. We have these avatars which is the sort of core of our product. Um And I think after this video is done, we of course see what the video looks like, but it's just to give you a quick visual of how easy it actually is to make videos with these systems. And it's one of the things that excites us the most about this new wave of generative tools and it's awesome if we can make it easier for Hollywood people and video production professionals to make video. But I think what's much, much more interesting is taking everyone in the world, including everyone here and making you into video creators and in some ways it sounds a bit science fiction in other ways. I think it's just natural evolution if you go back 4050 years in time and it wasn't part of most people's job to write texts, right? You had like secretaries, you had people in companies who would write things on typewriters and send them to the right people. Then we all got computers and keyboards and now I'm pretty sure everyone in this room you write as like part of your your daily tasks, right? Then powerpoint came along. We all became designers without really knowing probably all of you here know how to operate a powerpoint. And obviously, we think that the next iteration of this is going to be video, which is just a much, much better way of compressing information than text. At least for most people out there. I think with that, let's see the demo of what the final video looks like. This uses one of our latest technology, our latest model called Express One, which essentially teaches the avatars how they should perform and behave. But let's see the video. Imagine stepping into a room where the air crackled with the energy of innovation, a place where the future of A I and its impact on our world is not just discussed but shaped. Welcome to Fortune brainstorm A I in London. A gathering that promises to be the nexus of A I's brightest minds from leading technology companies at Fortune brainstorm A I in London. Expect to dive deep into conversations that matter. Hear from Google, deep minds, vice president and faculty A I CEO, as they unveil the future of A I and its transformative potential on society. Microsoft's chief scientist alongside accenture's chief A I officer will explore the profound changes. Generative A I is set to bring to the workplace. This event is not just about listening but engaging with roundtable sessions and ample opportunities for networking. I hope you enjoy the event. Great, very impressive given how quickly it all generates. Um Yeah, absolutely. Thank you. And now we'll hear from Maty and about 11 labs which is uh you may know about voice cloning. Let's see how this works. Thank you, Jeremy. I will attempt doing it live and maybe that's a very quick introduction. 11 labs is a Audi research and deployment company with the goal of making content university accessible across voices and languages. And what I will show you now is free demos, quick demos of how some of those technologies come to life. We should get a visual any second now to the screen. And the first of those building blocks is a one that most of you might be familiar with, with foundational text to speech. And that's one of the breakthroughs that we've, we've, we've figured out of how you can take existing text. And by understanding the nuance of the context, turn it into emotions and into the right intonation with a specific voice. So queuing up, I will use 11 apps to help me with the introduction to all of you and we'll see how that comes across. Ladies and gentlemen, welcome to the Fortune A I Brainstorm summit. We are thrilled to have you join us from around the globe for this dynamic gathering of minds and ideas today. Hm, we stand on the brink of new discoveries and innovations in the field of artificial intelligence that promise to redefine what's possible in business society and beyond, you could hear the pauses, you could hear the information, you could hear the excitement and also a little bit of ponder with the hm uh as, as as the A I was speaking and this is one of the voices from one of voice actors we work with. And actually there is plenty of those voice actors as part of our platform. We now have over thousands of voices which created clones of their voices and then shared it while every time it's being generated, they earn compensation and return as well. And whether you want this voice, a voice with an Australian accent with a different style or gender, this is all possible within the platform. But this is not again. And one of the things that's super exciting about the technology that it not only allows you to create the content in the language of the speaker, but actually take the content and turn it into other languages while preserving the same voice and the same characteristics. And I'll swap over to a demo that some of you might be familiar. It's a John F Kennedy speech from moon landing. And we'll see how initially he speaks in English and then I will flip it to Spanish and to Hindi life and you will see how some of those characteristics come across while we play the demo. But why some say the most, why choose this as our goal? And they may well ask why Kala Montana Masala Porque A Sua Atlantico hum for J for John. That's one small step for man. I didn't say I read it and you can see the emotions come across while the moon and the liftoff happened. And that's something we're excited about how you can take that content and enable those global audiences to watch it while still enjoying that original experience. But of course, in the audio world, there is so much more than speech and the voices. And as you think about the future about our work is how we can enable some of that work across some of the tangential domains, the sound effects, the music, how we can bring it all together with the video and truly make it an immersive experience and to close it off. I also show you something that you might be familiar from text to video work with Sora from Open Ayes as well where we'll take some of the stitches from the videos and we'll supplement it with A I generated sound effects. What we are trying to do is hit that unmute button with those videos that you've probably seen online out there. The real in a place beyond imagination where the horizon kisses the heavens. One man dares to journey where few have ventured armed with nothing but his wit and an unyielding spirit. He seeks the answers to mysteries that lie beyond the stars. And that's the goal to make all the all the content out there accessible across voices, across languages, across sounds. And uh and thank you. Well, I think you all agree. Those are pretty impressive demos. Um And you know, amazing technology also a little bit scary. Um And I think looking at this, a lot of people immediately think about uh some of the negative use cases around deep fakes, around fraud. Um I know there's already been some concern around 11 labs. Uh You know, whether people have used voice clones to uh perpetuate frauds. Uh how are you guys trying to, to prevent that from happening? Um And to make sure these technologies are not used uh for harmful uses. Uh Victor, maybe you go first actually. Yeah, sure. I think, I think any most new technologies, we immediately go to all things, things go wrong. Right. And I think in this case, that's definitely right. These technologies will be used by bad actors to do bad things. For sure. I think we should not be staring about that at all. And so for us, the the sort of safeguarding the technology has always been, has been been part of the company from day one. So we found the company in what we call our, our ethical framework, which is the free c consent, control and collaboration. Um It's a long topic, but I think that the kind of fundamental keystones for us is around consent. So never ever recreate anyone's voice or video avatar without the how do you, how do you guarantee consent? So we have a KYC style check when you go through, right? You submit your avatar footage. It is reviewed by a human being. You have to stay out of specific uh specific sentences to make sure you can make your clone. So essentially it is impossible today to go in and take some youtube videos or something and make a clone or someone that is just like not possible. Um And that we intend to keep it that way controls about content moderation. So we employ pretty heavy content moderation. There are quite strong stance on what you're allowed to create what you're not allowed to create. I think for most of us, the sort of what we call the red content is easy to agree on hate speech, violence, swearing, things like that. What gets harder is, well, someone's making a video about cryptocurrencies, for example. Right. Are you talking about what a great technological invention that is? Or you're trying to lead me into some schemes that promise I'm going to get rich in 10 days. Right. And we do our best to catch that stuff. It's a huge effort. It's of course a lot of computers and A I, but it's also a lot of humans sort of figuring out who are the bad faith actors on the platform. So that's a big, that is essentially internal product for us that said it is an incredibly small amount of users we catch trying to do this, but we do think it is our responsibility to, to take some stance on that. And the last one is Maddie. I want, I want to get you in there. How, how are you 11 labs are you trying to prevent this from misuse? Yeah, I think first of all, Jeremy, you're right, the defects are a serious concern, as we're saying about this year, air generated content and audio images videos coming coming across. Um And it's something that's a concern because open source other commercial models are becoming more active and, and there will be bad actors that will use and abuse them to the to the to the cases at 11 labs, it's education making everyone aware of the technology. Second is traceability. All the content that's generated by 11 apps can be traced back to the specific user specific account. And as they use the technology, they will need to verify themselves through different steps depending on the technology they use. And the last one is traceability, but also how you can embed the traceability as part of detection. So effectively all the content out there should be known as they are generated and there should be tools that allow you to quickly get that information as a viewer, then we release our tool publicly that allows everybody to use it and find whether it was 11 apps or not and reported back to us. So right now that that exists, if I run across some content and audio clip on the internet, on X or on linkedin or whatever I can check and see if it was 11 labs generated. Yes, you can, we can you can directly go on our website, upload that clip and and get probability of how likely it is. And now we are working with some partners to extend it across other technologies. So some of the open source model, other commercial models, that's part one and then part two because as you said, you can run into that in social media, but you might not be aware that it's A I generate in the first place. So how we can collaborate with social media, with telecommunication companies to check it on the fly while that content is being sent to you. I want to ask a question about the ethics of some of this. There's an interesting use case on the JFK speech. There have already been some cases of politicians in, in various countries wanting to appeal a certain audience in a language that they don't speak and they have translated their speeches in real time. I think maybe even using some of this technology to, to appear to speak those languages and they've said, oh, that's legitimate, I'm just trying to reach a different audience, you know, with my content, with my political content. But other people have said no, that's a complete red flag. You don't actually speak that language, you're presenting yourself with something you're not. What do you have a policy around this? And also at synthesia, do you have a policy around the creation of political content? Yeah, so we don't allow political content today. Um I think they'll change over time. Uh But I think we, we take a generally quite permissive stance on what you can use these technologies for. We also may be the enterprise. Um I do think it's a very interesting philosophical question. I think in five years time, something like that, no one will be talking about this anymore and everyone will be doing it. But I think as with any new technology, you will have this sort of years, right. Where people are trying to figure out what's right, what's wrong. Um, you could say, is it ok for a politician to have a ghostwriter to write a piece in fortune? Right. They didn't actually write it themselves. Is that worse? The same? Is it ok. Um, so I think, you know, all new technologies come with these sort of interesting questions and dilemmas and some of them will ring out to be very important and very, very true. And I think some of them will, will, will care less about in the future. And, and that idea of a policy definitely echo some of the vector pieces as well and where uh it's going to happen at one stage. And now as a society, we need to come together and figure out what's the right way of approaching that. Currently, we don't allow people around to impersonate any of that work or pretend that they are part of the candidates. Uh uh uh And the p and do the political speech or political content. Great. I want to take one question from the audience because I, I think I only have time for one. But if, if there's a question, please raise your hand and I'll try to come to you right here. I see as a gentleman here at this table, if you raise your hand again. Oh, great. If you could stand up and I yourself, uh it's Mark Salmon from the times, um uh when um Sadiq Khan had a deep fake of him, uh made the Met investigated and they concluded that they, they couldn't take any action. Um Do you think it's possible to write a law uh in this country to combat against that kind of deep fake? Um And do you think that's desirable, should there be laws against deep fakes? Obviously, I'm not a lawmaker. So I can't give you a definitive answer that I can give you my personal opinion. I think ultimately deep fakes is an advanced version of impersonation and impersonation is in most cases illegal if you do it for deceptive reasons, like fraud, for example, I definitely think we should include deep fakes of that, maybe even make the punishment harder if you're doing, using a deep fake for malicious impersonation. But where to draw the lines legally, I think it is very difficult to say that's not my, my my and to add uh one quick note to that as you, as you rightly ask, this content is out there, there's plethora of tools that can generate that content and what's the true solution? There's definitely the legislature that, that, that uh the countries can pass. But then also on the technical level, what we can do and beyond trace of and being able to detect it as you think here is out how this could work as part of the society. And one thing we are advocating for is beyond water marking. What's A I approved A I maybe there's a version of water marking what's real content and for Sadiq Khan or any other candidate to be able to transmit a message and that message is decoded. This is really Sadiq Khan that's speaking. So we don't only detect for what's fake but actually verify what's true. Excellent. Uh We're out of time. Thank you very much for the demos and thank you very much for the conversation. Uh Put your hands together, please for Maddie and.