TAll the preachers say The prediction that AI will take over the world comes from a famous line venture capitalist Marc Andreessen said about software, and in the past few weeks, it’s finally started to make sense.
This spring, tech companies made it clear that AI will become a defining feature of online life, whether people want it to or not. First, Meta wowed users with an AI chatbot that would reside in the search bar on Instagram and Facebook. It has since notified European users that their data is being used to train AI, presumably only to comply with the continent’s privacy laws. OpenAI released GPT-4o, touted as a new, more powerful and conversational version of its large-scale language model. (The launch event featured an AI voice named Sky that Scarlett Johansson claimed was based on her own voice without permission, but OpenAI CEO Sam Altman denies this. You can hear it for yourself here.) Around the same time, Google launched an “AI Overview” on its search engine, then scaled it back somewhat. OpenAI also reported that a number of media organizations ( Atlantic) and Reddit appear to operate under the assumption that AI products will become the primary way people receive information on the internet.AtlanticOur agreement with OpenAI is a corporate partnership. Atlantic Nvidia, which makes the microchips that power AI applications, reported record profits at the end of May and has since seen its market capitalization rise to more than $3 trillion. To sum up the moment, Nvidia’s $100 billion CEO, Jensen Huang, got rock-star treatment at an AI conference in Taipei this week, signing women’s breasts like a member of Mötley Crüe.
The pace of adoption has been dizzying, even unsettling, even to those who understand the technology best. Earlier this week, employees and former employees of OpenAI and Google published a letter declaring that “powerful financial incentives” are forcing the industry to avoid meaningful oversight. The same incentives also seem to lead companies to produce a lot of garbage. Chatbot hardware products from companies like Humane and Rabbit were touted as an attempt to usurp the smartphone, but shipped barely functional. Google’s hasty launch of AI Overviews produced ludicrously flawed and potentially dangerous search results in an attempt to compete with Microsoft, Perplexity, and OpenAI.
In other words, technology companies are racing to grab money and market share before their competitors do, and as a result, they’re making self-destructive mistakes. But even as technology companies create the hype train, other companies are happy to jump on board. Leaders across industries, terrified of missing out on the next big thing, are signing checks and signing deals, perhaps not knowing exactly what they’re getting themselves into, or whether they’re unwittingly helping the companies that will ultimately destroy them. The Washington PostVineet Khosla, chief technology officer at , has reportedly told staff he plans to put AI everywhere in the newsroom, even though its value to journalism is, in my eyes, unproven and merely decorative. We watch as planes are assembled haphazardly in the air.
As an employee of a publishing company that recently signed a deal with OpenAI, I have some insight into what happens when generative AI sets its greedy eye on a small corner of our industry. What would it be like if AI were to eat the world? It would feel like we were trapped.
THere are the elements These media partnerships feel like blackmail. Tech companies have trained their large language models with impunity, arguing that scraping internet content to develop their own programs is fair use. It’s the logical end point of Silicon Valley’s classic “ask for forgiveness, don’t ask for permission” growth strategy. A cynical interpretation of these partnerships is that media companies have two choices: take the money offered or allow OpenAI to scrape their data anyway. These terms are more like hostage negotiations than a mutually agreeable business partnership. This is what media executives are saying privately and sometimes publicly to each other.
The publishing industry can of course refuse such a deal. There are other options, but the technical term is: not goodYou could also sue OpenAI and Microsoft for copyright infringement. The New York Times As they did, they hope to establish a legal precedent that exploitative generative AI companies will be fairly compensated for the work used to train their models. This process would be prohibitively expensive for many organizations, and if they lose, they will only get legal fees. That leaves a third option: to eschew the generative AI revolution entirely in principle, block the web crawling bots of companies like OpenAI, and take the rightful moral stand while their competitors cave in and take their money. This third path requires betting on the hope that the age of generative AI is over-hyped. Times It’s unclear whether the lawsuit will be successful or whether the government will step in to regulate this exploitative business model.
The situation facing publishers seems to perfectly illustrate a broader dynamic: no one knows exactly what to do. This is unsurprising, given that generative AI is a technology that has been defined by ambiguity and lack of consistency thus far. A Google user encountering AI Overview for the first time may not understand what it’s for, or whether it’s more useful than regular search results. There’s also a gap between the existing tools and the future we’re being sold. We’re told that the innovation curve will be exponential. We’re warned that the paradigm is about to shift. We’re forced to believe that ordinary people have little choice, especially as computers get larger and more powerful. We can only experience mild disorientation as we shadowbox with the notion of this promised future. Meanwhile, the ChatGPTs of the world are being forced upon us by tech companies that insist that these tools should somehow help us.
But there’s another side to these media alliances, and they suggest a moment of cautious opportunity for beleaguered media organizations. Publishers already Algorithm suppliers and media companies have been treated unfairly for decades, allowing platforms like Google to index their sites and in return receiving only traffic referrals. By this logic, signing a deal with OpenAI is not a surrender or good business, but a way to fight back against the platforms and set the ground rules. You have to pay for our content, and if you don’t, we’re going to sue you.
After speaking with several executives from various companies in negotiations with OpenAI over the past week, my impression is that the tech company is more interested in real-time access to news sites for its upcoming search tools than it is in publisher data to use to train its models. (I agreed to keep these executives anonymous so they could speak freely about their companies’ deals.) Having access to publisher partner data is useful to the tech company in two ways. First, it allows OpenAI to cite third-party organizations when users ask about sensitive issues, so it can claim it doesn’t make editorial decisions about its own products. Second, the company needs up-to-date information if it has ambitions to overtake Google as the dominant search engine.
Here’s where we hear media organizations could have an advantage in the ongoing negotiations. OpenAI would, in theory, still ask for up-to-date news information. Other search engines and AI companies that want to compete would also need that information, but now there’s a precedent that they should pay for it. This could create a steady revenue stream for publishers through licensing, which is unprecedented. Record companies have fought platforms like YouTube over copyright issues and found ways to get paid for their content. That said, it’s not like news organizations are selling Taylor Swift songs. (OpenAI and AI spokespeople say Atlantic What was revealed to me was AtlanticThe two-year contract allows the company to sell its products Atlantic Ocean However, if the agreement is terminated, OpenAI will no longer be permitted to use the content unless it is renewed. Atlantic Ocean Data for training a new underlying model.
figureout But even this optimism is beset with problems. Do we really want to live in a world where generative AI companies have greater control over the flow of information online? The shift from search engines to chatbots would be highly disruptive. Google is imperfect, and its products arguably underperforming, but it has provided a foundational business model for online creative work by delivering optimized content to audiences. The search paradigm needs to change, and it’s only natural that the web page will become a relic. And yet, the scale of the disruption, and the thoughtlessness of tech companies to suggest everyone join in, gives the impression that no AI developer is interested in finding a sustainable model for creative work to thrive. As Judith Donath and Bruce Schneier recently wrote in this publication, AI “threatens to destroy the complex online ecosystem that enables writers, artists, and other creators to reach human audiences.” Following this logic, things quickly become existential. What incentive would people have to create work if they couldn’t make a living from it?
If you feel your brain starting to spin in your head, you’re in for a full experience of the generative AI revolution sweeping through your industry. This is disruption. actually It feels chaotic. It feels hectic. People tell me it’s an exhilarating moment full of opportunity, but I’m not really sure what that means in practice.
No one knows what comes next. Generative AI companies have developed tools that nominally help improve productivity but are popular and only a faint shadow of the ultimate goal of building human-level intelligence. But these companies are extremely well-funded and aggressive, and they can leverage breathless hype cycles to build muscle and charge headfirst into any industry they choose with the express purpose of making themselves a central player. Is this moment’s technological advancement worth the disruption? Or will the hype slowly fade, leaving the internet even more broken than it is now? After a couple of years of the latest wave of AI hype, what’s clear is that these companies don’t need to build Skynet to be disruptive.
AI is eating the world Proponents of technology mean this as a triumphant, exciting phrase. But that’s not the only way it can be interpreted. It can also be read threateningly, as a battle cry for rapid and aggressive colonization. Lately, I’ve heard it with a shrug of shoulders and a tone of folded hands and resignation. What’s left unsaid is what happens after the raw food is consumed, digested, and the nutrients extracted. We don’t say it aloud, but we know how it goes.