
Generative AI Is About To Reset Everything, And, Yes It Will Change Your Life | Forbes
What this covers
This Forbes video is a reported explainer on the generative AI boom of late 2022 and early 2023. It mixes a journalist's narration with commentary from an AI-governance voice, and the speakers are not named in the transcript. It begins with the field's recent history: OpenAI started as a non-profit backed by Sam Altman, Elon Musk and Peter Thiel with the aim of building artificial general intelligence (AGI)). The video then follows the viral text-to-image releases of summer 2022, including DALL-E and Stable Diffusion from Stability AI, and the November launch of ChatGPT. It also covers Microsoft's roughly $10 billion investment in OpenAI and Bill Gates's ranking of the technology alongside the PC, graphical interfaces and the internet.
The middle stretch is about use cases. These include search and moderation, video backgrounds, image, text and 3D generation, translation, homework help, and automated communication. The journalist argues that human correspondence is more predictable than people assume, and that he is not worried about being replaced. A newsroom example raises a separate concern: some outlets publish AI-written articles without disclosing it.
The problems are the most contentious part. Voice cloning helps with translation but also enables forgery. Models "hallucinate" confident wrong answers. Getty's suit against Stability AI shows that copyright is unsettled. The guest says responsibility standards are slipping as cheaper entrants drop safeguards. The guest also looks back at the earlier wave of ranking systems at Google, Facebook and YouTube, whose problems remain unresolved. On governance, the guest rejects both unlimited individual power and corporate control, and proposes Platform Democracy, in which representative global populations deliberate on deployment. The guest also warns of Reality Apathy, where people stop trusting that anything is real. The video ends by saying how AI was used to make it.
Generative AI represents a transformative second wave of AI that will reshape society through content creation rather than ranking, but requires democratic governance frameworks and responsible development practices to avoid harms like deepfakes, copyright violations, and erosion of shared reality.
- First wave (ranking systems like Google/YouTube) reshaped incentives; second wave (content generation) will be far more consequential
- Current governance approaches—pure individual power or tech company control—are both inadequate; democratic processes needed at global scale
- Without responsibility safeguards, generative AI enables voice cloning, deepfakes, and misinformation at scale, eroding societal trust
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Generative AI enables voice cloning technology that is useful for translation (enabling people to speak in their native language with AI preserving their voice), but the same technology enables malicious voice forgery where someone can fabricate audio of another person saying anything without consent.
“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”
AI systems are prone to 'hallucination'—generating confident but completely incorrect answers—exemplified by ChatGPT confidently providing wrong answers to basic math problems, which could harm businesses relying on inaccurate AI-generated information.
“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”
AI can improve existing use cases like search and moderation in social networks, improve communication by adding virtual backgrounds to video, and create completely new use cases that were impossible before like generating images, text, 3D content, voice recognition and translation, and conversational AI.
“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”
OpenAI launched ChatGPT on November 30th as a tool that enables users to request content in specific styles (e.g., a recipe in Shakespearean form, scripts in the style of movie stars), which demonstrated to the public that AI was breaking into the mainstream beyond just tech industry interest.
“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”
OpenAI started as a non-profit backed by Sam Altman (former Y Combinator president), Elon Musk, and Peter Thiel with the goal of creating artificial general intelligence (AGI)—a sci-fi-level AI that would be conscious, self-learning, and capable of rethinking capitalism.
“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”
Copyright and intellectual property issues remain unresolved in generative AI: models are trained on massive internet-sourced datasets without compensating creators, as exemplified by Getty Images suing Stability AI for training on millions of Getty photos without payment.
“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”
Public understanding of AI capabilities and limitations is critical; everyone needs to understand what AI can and cannot do, including its vulnerabilities (like bias), so society can collectively mitigate risks and avoid unintended consequences.
“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”
Generative AI will impact virtually every domain where humans interface with computers, from chatbots to media production to education and healthcare, making comprehensive change difficult to anticipate.
“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”
The reason platform/AI democracy needs to be global is because the technology affects everyone globally, and this provides legitimacy that makes it hard for companies to claim they're democratizing technology when they're actually just giving everyone the power to harm things.
“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”
Bill Gates considers generative AI one of four major technological changes in his lifetime, alongside the PC, graphical interfaces, and the internet, and believes it will deservedly be the hottest topic of the year.
“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”
The current generative AI wave has been building for several months, reaching an inflection point around summer 2022, with companies like OpenAI (DALL-E 2), Stability AI (Stable Diffusion), and others releasing text-to-image models that achieved viral adoption and mainstream breakthrough.
“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”
The first wave of transformative AI was ranking systems (Google search, Facebook, YouTube recommendations, TikTok) that chose which already-created content to surface to users, which restructured societal incentives for politicians, entertainers, and journalists despite seeming like a simple technical task.
“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”
Some media companies are beginning to use AI tools to generate articles, with some companies publishing AI-generated content without disclosure, raising ethical and transparency concerns about AI-generated vs human-created journalism.
“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”
There is real risk that society will move toward 'reality apathy'—where people stop believing anything is real and give up on distinguishing truth from fabrication—which would destroy functional society and democracy because both depend on trust and shared understanding of facts.
“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”
AI systems should optimize for human values that support society and democracy (not just engagement), and critically, should not optimize for division—which is the natural algorithmic output of engagement-based optimization—but instead should optimize for bridging divides and making conflict productive.
“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”
A third governance approach exists: bringing technology impact questions to democratic deliberation with representative global populations, asking how technology should be deployed and creating legitimacy through inclusive decision-making processes.
“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”
Two main governance camps exist: one giving individuals unlimited power to use AI, the other giving tech companies control over what users can do, but both approaches have serious downsides because individual power without constraints enables harassment, extortion, scams, and democracy manipulation, while corporate control removes individual agency.
“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”
Microsoft has invested approximately $10 billion in OpenAI at a $29 billion valuation, becoming a key allied partner.
“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”
Generative AI is being applied in education where students are using it to help with homework structuring and writing, expanding its impact beyond creativity into learning support.
“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”
Some AI companies have built safeguards to make voice cloning difficult without consent, but responsibility standards are collapsing as new entrants undercut these practices because it's cheaper and easier to launch without safety features, lowering the industry's responsibility baseline.
“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”
The rapid rate of AI advancement over the next few years will outpace societal preparedness, creating risks because existing problems from the previous wave of AI (ranking systems) remain unresolved while new generative AI problems accumulate.
“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”
AI systems are good at predicting communication patterns (email rejection patterns, PR pitch styles, reporter response preferences), suggesting that human communication is more predictable than commonly believed and AI will be useful in automating communication earlier than expected.
“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”
The journalist is not concerned about AI replacing their work because if AI can replace their job, they are not performing journalism well enough; instead, they will use AI tools like other technologies (iPhone, transcription) to do their work better.
“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”
Unlike other technologies like the metaverse or web3 that always created their own systems without immediate value, generative AI can integrate into existing systems, making it more likely to succeed and provide immediate value to organizations.
“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”
Previous technological waves (machine calculation, information retrieval via search engines) enabled new use cases that were previously impossible, and the same pattern is occurring with machine learning and AI, making current hype sustainable rather than speculative.
“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”
Big tech companies (Microsoft, Google) will integrate generative AI models deeply into existing software (Word, Outlook, Google Docs) rather than AI-native applications replacing them, creating an open question about whether entirely new AI-native software will emerge or existing platforms will dominate.
“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”
The coolest and most transformative applications of generative AI have not yet been invented or even imagined, suggesting most value creation is still ahead and initial applications (business workflow integration) are boring compared to future possibilities.
“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”
Machine learning represents a fundamentally different paradigm from traditional software, which will enable a new wave of AI/ML-native startups that can build products 100x better than incumbents who have not adapted to machine learning methods.
“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”
The current consensus on AI has shifted from pessimism about AI as a technology to cautious optimism, with people now accepting AI as inevitable and debating how to manage it responsibly rather than whether it should exist.
“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”
Stability AI's competitive advantage is that they are open source and release their models, building an amazing community that outstripped Bitcoin and Ethereum code popularity in just a few months, compared to those technologies' 10 years.
“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”
Generative AI is like a 'rocket ship for the mind'—significantly more powerful and transformative than the smartphone (which Steve Jobs called a 'bicycle for the mind')—in enabling humans to externalize and manifest their thoughts.
“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”
Bill Gates is optimistic about AI's impact in developing countries, particularly in Africa where AI-based medical question-answering tools could address doctor shortages without replacing physicians, and in low-income US communities where AI tutoring could supplement limited access to human tutors.
“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”
Generative AI works by taking a large dataset, receiving a user prompt, and outputting something completely new that has learned from that data to answer the prompt.
“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”
OpenAI declined to be photographed for Forbes' AI coverage, so the magazine used DALL-E to generate an image of Sam Altman, producing weird results including recursive uncanny-valley images of a blurry face holding a magazine with the same blurry image.
“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”