Generative AI Is About To Reset Everything, And, Yes It Will Change Your Life | Forbes

generative AI is in a sense the second wave of transformative AI that we've experienced over the past few decades and that first wave was really around ranking systems and so you're thinking here about Google search Facebook YouTube's recommendations Tick Tock all of this is around choosing sort of the ranking of information that was already created by people and this seems like a small thing it's just ranking a set of items but when you're ranking everything that Humanity has created that actually affects the incentives of politicians of entertainers of journalists you're now restructuring Society just by that simple Act of ranking because you're ranking for billions of people every day all the time the second wave is where we go beyond ranking to actually generating that content this second wave of transformative artificial intelligence we don't know fully what those implications are going to be the whole way in which we make information and the basis of communication across Society like that is all going to be rewired [Music] at its simplest generative AI refers to an AI tool that is generating or creating content that could be text audio or video you're taking a really large data set and you're inputting a prompt and then you're outputting something completely new that is learning from all that info and creating something to answer the prompts we've covered AI for some time at Forbes but on November 30th when openai launched chat GPT it's tool that allows you to you know ask for a recipe or a sauna in the form of Shakespeare or a script in the form of you know Alex as movie star which my boss did he also asked it to write an annual performance review for me which I did not appreciate well we saw everyone playing with these tools we thought okay AI is really breaking into the mainstream here it's not just something that the tech bubble cares about and it's time to ask questions about who's really gonna make money here how real is this hype and what will be the business impacts openai is a really interesting company it actually started as a non-profit it was backed by Sam aldman who used to run my combinator as well as the billionaires Elon Musk and Peter Thiel and it was formed with the idea of creating an artificial general intelligence this is a real kind of sci-fi AI one that is conscious self-learning that could theoretically rethink capitalism as Sam almond has said or could potentially outstrip human Control building that AGI was something that openai could not do as a non-profit it couldn't raise enough money to do this so it needed to create a for-profit entity openai does that in 2019 and that's when Sam Allman becomes CEO full-time and not long after we've seen the company become a key Allied to Microsoft which has reportedly invested 10 billion dollars recently at a 29 billion dollar valuation foreign this current wave that we're seeing this generative AI wave has been kind of building up for the last couple months now really hitting an inflection point probably summer of 2022 we were slowly seeing these different companies starting to tease out you know new releases like open AI launching Dolly 2 and beta and then stability AI announced stable diffusion this text to image generation model I think really it was the culmination of all these different models that reached a viral moment you know really got to the masses I would characterize the consensus as cautiously optimistic you know I think there's not so much pessimism around like AI as a technology anymore I think more than like optimism it's kind of gotten to a point where people are accepting that this is an inevitability now what do we do with this inevitability you know different folks have different levels of bullishness on where this technology can take us or what the potential harms and risks of it are but uh for the most part people seem to be optimistic that we still have time it's still kind of in our hands to kind of control the destiny that this new AI Paradigm Shift brings I think the way to look at this is if Steve Jobs said that you know the phone was a bicycle for the mind this technology is really a rocket ship for the mind the most exciting application I think of this generative AI is with what we're building allowing humans to communicate visually so you can take what's in your mind and create it almost instantly whereas before it would have taken days or months even for skilled individuals to do that via art or PowerPoint or whatever else I think we've also seen the impact of this also emerging other sectors such as education where students are now using it to help with their homework structuring or writing I think we'll see it across everything whereby again anything you can imagine you'll be able to create now stability or differential is that we are open source so we release our models so there's an amazing Community that's built around it if you look at popularity on the code base it's outstripped things like Bitcoin in ethereum in a few months what took them 10 years the types of applications that we're seeing are largely around basic creativity at the moment but now they are extending into combining these things to create video game assets to create artwork of various types to create new textures and things like that to extend the capabilities of creators and to make more people creators I think where it goes from here I think we'll see more and more business workflow use cases definitely expect to see these big tech companies like Microsoft and Google that are building some of these models and kind of working with companies like open AI on these models to start integrating them more deeply into some of the existing software we know so for example you could expect to see generative AI use cases in Microsoft Word or Microsoft Outlook Google Docs stuff like that I think one of the open questions remains are we going to see like an entirely new wave of uh AI native applications kind of become a whole new gen generation of software tools that people use or will it be kind of the existing options that are able to integrate this AI faster and stay ahead of the curve for example Adobe putting this AI into Photoshop versus entirely new like AI powered photo editor coming about speaking of Bill Gates about AI I was honestly a little surprised just how excited he was I asked him how does this compare to previous technological Cycles like the the rise of the PC or the PC with you know Graphics interfaces that allow us to play games and do a lot more you know graphically inclined things on computers or even the internet itself and Bill said that he thinks this is one of the four big technological changes of his lifetime and he said that he thinks AI will deservedly be the hottest topic of this year overall I would say Bill Gates was optimistic about the impact that these AI tools could have particularly in the developing world where he spends a lot of his thoughts with his foundation he was thinking about how in Africa some people may not have access to human doctors you can do a lot as a foundation to try to get doctors to that place or train new doctors but an AI tool that could answer some basic medical questions could fill a need without challenging doctors professions similarly he was thinking in you know lower income neighborhoods here at home in the U.S there might be some math tutoring or basic homework support or teaching support but a tool like this could use for students who don't have access to human Tutors or don't have a lot of time with a teacher so with AI you can do existing use cases better than before right so you can improve search for example you can improve moderation in social networks you can improve communication by adding uh virtual backgrounds for example to to this video you can also create completely new use cases that were just not possible before right you can generate image Club text you can generate 3D context you can like recognize voice and translate that you can create better conversational AI so it really opens up a new range of possibilities the same way the previous generation of Technology creates the ability for machines to calculate better or to retrieve information better with search engine there's a whole range of new use cases that weren't possible before that are now made possible by Machine learning uh the hype that we're seeing now is just to catch up on the previous wave of usage this is very different for example than crypto where you had a lot of pipe but very little usage I think it's quite sustainable and as we see more use cases more usage more companies investing on the topic in 2023 hopefully it's gonna going to become more and more for mainstream topic and more and more Central topic what we want is for everyone to understand AI understand what it can do what it can't do some of the challenges associated with it for example biases are a big challenge for AI systems today we want everyone to you know understand that so that they can you know help us mitigate some of the risk uh and and make sure it doesn't create more challenges so there are a number of issues that fall into different categories that still need to be resolved with this AI as uh flashy and exciting as some of this new technology is it's still prone to errors and mistakes AI practitioners call it like hallucination where you might ask open AI is a chat gbt a basic math problem and it will very confidently give you an answer but it turns out the answer is completely wrong that could you know in a business capacity hurt a business if they're kind of using inaccurate information taking it at face value beyond that there are kind of more deeper ethical concerns that have not been addressed one of them being the copyright and IP around this new AI the way that this generative AI is possible is these companies have trained their AI models on this massive massive quantities of data and a lot of that data is pulled from the internet you know so take an example like stable diffusion which is trained on millions and millions and millions of images around the internet stability AI which is one of the companies that we write about in in our coverage and is behind stable diffusion they've been sued recently by Getty Images which is alleging that they used millions of Getty's photos to allegedly train the models without paying Getty a dime for their work there's also a lawsuit going on against openai Microsoft and GitHub which is owned by Microsoft from programmers who believe that the code that you can get which is usable it's not perfect but it can actually save time for some programmers today but that code comes from programmers who similarly were not compensated for their work foreign [Music] the rate of change that I expect that we're going to be seeing over the next few years it makes me concerned that we are not going to be ready for what's coming next and yes we're better at the issues from 10 years ago but we still haven't even solved the problems that have been caused by the the old transformation where ranking systems sort of reshape our society right now there's two main sort of camps within the tech World in terms of how we're going to govern this technology and one is we're just gonna give everyone the power as an individual to do whatever they want with it and the other is that the tech companies themselves will decide what you can do and both of those don't seem really great like yes it's nice for us to have individual power but if everyone has the power to mess with democracy everyone has the power to harass others to extort others to generate scams quickly on demand like that maybe isn't really good for the world there's a question of what else is there right and then that is what democracy tries to do it's it's can we take that question of how this technology should impact society and bring it to a process where those decisions are being really considered in terms of their impacts but the challenge is that our current Democratic systems are frankly too slow and the AI advances are both incredibly quick and they're global one of the most terrifying and in many ways exciting applications of generative AI is to copy our voices that's incredibly useful for translating across languages so you can have you just speaking in any language and it really sounds like you but the downside there is someone can use that same technology to just copy your voice saying anything they want we had companies in the space that were really responsible as well as one couldn't be and they they had a lot of safeguards to make it so that you really it was really hard to copy someone's voice about their consent but that all changed very recently when a number of companies just came out without doing that it's sort of like you're lowering the bar for responsibility because these new entrants are coming in because it's cheaper to not build in that extra Tech and it makes it easier to use and so we can't live in a world where people keep undercutting those responsibility practices there is real risk that we are moving toward a world where people just don't have any sense of what is real and they just sort of give up on that and we've already seen that obviously in many parts of the world the sort of what I'm calling reality apathy that is not the ingredient that we need if we want to have functional Society functional democracy like Society is built on trust economics is built on on trust politics is built on trust and we get into a really dark place when we don't know what we can believe I think there it's very difficult to see where this AI will not impact anywhere that a human and a computer interface from better chat Bots to the whole of the media industry changing to create faster better output and even areas like education and Healthcare augmenting human potential in these areas it's hard keeping up to date with what giant Fai look likes next week let alone one year or five years but I think in the next few years we're going to see video we're going to see audio 3D the ability to really create any world you can imagine and I think that will be amazing I think we'll see it in areas like education and Healthcare and others really helping and enhancing our ability to Aid people and I think that's going to be great to see because it can integrate into existing systems unlike other things like maybe the metaverse or web 3 that always created their own systems and there was never really immediate value for a couple things that have shocked me about AI as I've gotten to know this space one is that I feel like we're not all as unique as we think we are these tools are pretty good at predicting what an email to reject a PR pitch looks like or right back at me you know the kind of way to send a PR pitch to someone like Alex the reporter that kind of communication is actually pretty well predicted and so I think that we're going to see these AI tools you know being useful there much earlier than I would have guessed Sam Walden declined to be photographed for our Forbes story we did photograph Greg Brockman and a couple other leading entrepreneurs in this space but for Sam we thought it would be really fitting to use his own tools to generate an image of himself so he put into a prompt in open ai's image generating tool called dolly that said Sam Alden on the cover of Forbes we got some weird ones back including a cover of sort of a blurry faced uncanny valley Sam almond-like person holding a cover of a magazine like Forbes with the same image of blurry Sam Altman we've already seen some media companies starting to explore how to use these AI tools themselves some elegantly some actually getting in trouble for publishing stories written by an AI tool and not disclosing it right away I think this is a debate that's going to happen across the media landscape for someone like me I'm actually not concerned I do believe that just like using an iPhone or a transcription tool or other pieces of technology in my job I'll be using AI to do my work better if the AI can replace my work then I don't think I'm doing a good job thank you I think her sense talking to these Founders these investors is that uh a lot of the coolest applications that are going to come out of this you know AI wave haven't even been invented yet maybe haven't even been imagined yet that said it seems like for the most part starting out where this AI is going to be applied is in some of the more boring places some of the more busy workplaces how do we integrate this AI into our basic business workflow so I think you're starting to see kind of consumer behaviors change students for example are already kind of using chat gbt to generate essays for them some artists are using models like stable diffusion mid-journey or open ai's Dolly to help them create works of art I think because machine learning is such a different Paradigm as a typical software there's going to be like a flock like a wave of AI ml native startups that are going to do things build products 100 times better than the inconvenience who have not been able to deal with machine learning the the same way so this is very exciting because I think we really have the opportunity to change the Paradigm completely and make a merge a whole group of newcomers uh and and new companies that are going to be able to do things differently what we can do is you can use sort of new forms of democratic processes and this is what what I've been calling sort of platform democracy or AI democracy to ask a sort of Representative selection of the global population like how should this technology be used how should it be deployed it sounds crazy to to say this but there already are pilots of of this sort of global process and the reason why you need it to be Global is because this technology affects everyone it creates legitimacy it makes it really really hard for companies to then be like oh we're democratizing this technology when really what they mean is we're giving everyone the ability to just like mess with everything that isn't what a democratic process that isn't what considered deliberation would have led to if you had the people who were being impacted by that technology in the room we need our AI systems to not just be optimizing for engagement as we've seen with things like Facebook and Tick Tock and YouTube but for the sort of the human values that actually support society and support democracy and a big part of that is sort of not optimizing for division which is what comes naturally out of those algorithms in many cases and so a lot of my work recently has been around how can we instead optimize for bridging divides making conflict productive as opposed to destructive so that I think that is absolutely core to bake in from the beginning because we've seen what happens to society when systems that billions of people are using every day are really optimizing for division and it isn't pretty in my opinion Ai and machine learning is becoming the new paradigm to build all technology in my opinion Ai and machine learning is becoming the new paradigm to build old technology

Channel: Forbes

Views: 121,507

Keywords: Forbes, Forbes Media, Forbes Magazine, Forbes Digital, Business, Entrepreneurship, Technology, Investing, Personal Finance, Open AI, Generative AI, Stability AI, Hugging Face, Stable Diffusion, Sam Altman, Kenrick Cai, Alex Konrad, Aviv Ovadya, Clem Delangue, Emad Mostaque, Dall-e, democracy, fake news, misinformation, deepfake, bill gates, what is chatgpt, chatgpt, generative AI, billionaires, billionaire, future of AI, what is AI, how to use AI, is AI good, is AI bad

Id: WY518YRfs5M

Length: 21min 25sec (1285 seconds)

Published: Fri Feb 17 2023