For almost two years now, the world’s biggest tech companies have been at war with generative AI. Meta may be known for social media, Google for search, and Amazon for online shopping, but since the release of ChatGPT, each has made huge investments in an attempt to dominate this new era. Alongside start-ups such as OpenAI, Anthropic and Perplexity, their spending on data centers and chatbots is poised to eclipse the costs of sending the first astronauts to the Moon.
To succeed, these companies will need to do more than create the “smartest” software: they will need people to use and return to their products. Everyone wants to be Facebook and no one wants to be Friendster. To that end, the best technology strategy hasn’t changed: create an ecosystem that users can’t help but live in. Billions of people use Google Search every day, which is why Google created a generative AI product known as “AI Insights.” directly in the results page, giving it an immediate advantage over its competitors.
This is why a recent proposal of the Department of Justice is so important. The government wants to break Google’s monopoly on the search market, but the proposed solutions could actually do more to shape the future of AI. Google has 15 products that each serve at least half a billion people and businesses: a vast ecosystem of gadgets, search and advertising, personal apps, and enterprise software. An AI assistant that appears in (or works well with) these products will be the one these people are most likely to use. And Google has already integrated its flagship Gemini AI models into Search, Gmail, Maps, Android, Chrome, the Play Store and YouTube, all of which have at least 2 billion users each. AI doesn’t need to be life-changing to succeed; it just has to be frictionless. The DOJ now has an opportunity to add some resistance. (In a statement Last week, Kent Walker, Google’s chief legal officer, called the Justice Department’s proposed solution “an interventionist program that would harm Americans and their global technological leadership,” including America’s “leading role.” (company in AI.)
Google isn’t the only competitor with an ecosystem advantage. Apple is bringing its Apple Intelligence suite to eligible iPhones, iPads and Macs. Meta, with more than 3 billion users across its platforms including Facebook, Instagram and WhatsApp, enjoys similar benefits. Amazon’s AI shopping assistant, Rufus, has received little attention, but it has nevertheless become available to US buyers from the website this fall. No matter how much of the DOJ’s request the court ultimately grants, these giants will still lead the AI race, but Google had the clearest advantage among them.
The quality of these companies’ AI products is of limited relevance to their adoption. Google’s AI tools have repeatedly shown major flaws, like confident recommendation eat stones for good health, but the features continue to be used by more and more people simply because they are there. Likewise, Apple’s AI models are less powerful than Gemini or ChatGPT, but they will have a huge user base simply because of the popularity of the iPhone. Meta’s AI models may not be cutting-edge, but that doesn’t matter to the billions of Facebook, Instagram, and WhatsApp users who just want to ask a dumb question to a chatbot or generate a random illustration. Technology companies without such an ecosystem are well aware of their disadvantage: OpenAI, for example, is would have plans to develop its own web browser and has partnered with Apple to bring ChatGPT to the company’s phones, tablets and computers.
This is why it is relevant that the antitrust remedy proposed by the DOJ targets the broader Google ecosystem. Federal and state prosecutors have asked the court to force Google to sell its Chrome browser; stop prioritizing its search products in the Android mobile operating system; prevent it from paying other companies, including Apple and Samsung, to make Google the default search engine; and allow competitors to syndicate Google’s search results and use its search index to create their own products. All of these and other DOJ requests under the auspices of search are actually attacks on Google’s vast empire.
Like my colleague Ian Bogost has arguedChrome sale may not affect Google’s search dominance: “People came back to Google because they wanted to, not just because the company put pressure on them,” he said . wrote last week. But selling Chrome and potentially Android, while preventing Google from making its search engine the default option for various other companies’ products, would make it harder for Google to funnel billions of people to the rest of its software, including including AI. Meanwhile, access to Google’s search index could give a huge boost to OpenAI, Perplexity, Microsoft and other AI search competitors: maybe the hardest Part of building a search bot is scouring the web for reliable links, and competitors would have access to the most coveted way of doing so.
The Justice Department appears to recognize that the war on AI involves and goes beyond research. Without intervention, Google’s search monopoly could also give it an unfair advantage over AI, and an AI monopoly could further strengthen the company’s control over search. The court, the lawyers wrote, must prevent Google from “manipulating the development and deployment of new technologies,” including AI, to further strangle competition.
The order therefore also explicitly targets AI. The DOJ wants to ban Google from self-preferred AI products, in addition to search, in Chrome, Android and all its other products. He wants to prevent Google from buying exclusive rights to AI training data sources and prohibit Google from investing in AI start-ups and competitors that are or could enter the search market . (Two days after the DOJ released its proposal, Amazon invested another $4 billion in Anthropic, the startup and OpenAI rival that Google has also heavily backed so far, suggesting that the e-commerce giant could try to maintain an advantage over Google.) The DOJ also requested that Google provide an easy way for publishers to opt out of having their content used to train Google’s AI models or be cited in Google-enhanced search products. AI. All of this will make it harder for Google to train and commercialize future AI models, and easier for its competitors to do the same.
When the DOJ first sued Google, in 2020, it was concerned about the Internet of old: a web that seemed intractably blocked, long calcified in the image of the company that controls how billions people access and browse it. Four years and a historic victory later, the proposed remedy enters an Internet undergoing a upheaval that few could have predicted – but which the DOJ lawsuit nevertheless appears to have anticipated. A frequently raised issue in antitrust litigation in the technology sector is anachronism: by the time a monopoly in social media, personal computing, or e-commerce emerges, it is already too late to disrupt it. Thanks to generative AI, the government could finally have the head start it needs.