On a rainy Sunday afternoon in Delhi, 22-year-old Aditi is working on her college assignment about Climate Change, and the global music industry’s responses to it. Instead of the age-old, well, decades-old, habit of typing keywords into Google search, and sifting through dozens of links that it throws up, she turns to ChatGPT, the Artificial Intelligence (AI) tool that has become common among researchers, writers and common Internet searchers.
“Give me the latest on how the musicians are responding to Climate Change during their tours, and in their albums and productions,” she types. Within seconds, a neat, highly-referenced summary appears, which includes recent initiatives, industry examples, and even the mention of carbon-neutral concerts by the iconic band, Coldplay. Aditi’s example shows how the days of Google searches, opening multiple tabs, seemingly-endless clicking to read more and closing ads to get information are gone. Searches today are about AI-generated cogent and coherent summaries within seconds, and almost-seamlessly.
Google’s answer to this AI-shift is Gemini-powered AI Overviews, which do not generate contextual highlighted links, but scan multiple sources, weigh context, and stitch them together into a single, seemingly-authoritative response at the top of the results page. To most Gen Z and digital-first users, they look like a shortcut. For the tech giant, Google, they imply something bigger. A new search model to disrupt other AI-driven ones, which have become popular.
Gemini aids Google to counter the growing might of the likes of ChatGPT and others, which keeps the traffic intact, and keeps the users within its ecosystem for a longer period, which reduces the chances of the searchers wandering off to actual content and publisher sites through links. Both protect Google’s eyeballs, and advertising meat.
For publishers, who are dependent on Google for reach, and who painstakingly spend huge amounts to generate exclusive and relevant content, both ChatGPT and Gemini are dangerous, as they kill potential and actual revenues. Both models threaten their existence. Hence, independent content generators have taken AI-driven search companies to the courts, both in India and other countries.
Penske Media Corporation (parent company of Rolling Stone, Billboard and Variety) sued Google in a Washington court. The former alleged that its AI Overviews strip journalism for summaries, cutting publishers out of the traffic and revenue chain. Others have dragged ChatGPT into similar legalese. The lawsuits highlight an existential question: In a world where millions of users like Aditi increasingly seek answers (summaries), rather than links, how will the existing economics of information survive, and work in the future?
Larry Page, Google’s co-founder, once mused, “Artificial Intelligence would be the ultimate version of Google. The ultimate search engine that would understand everything on the web. It would understand exactly what you wanted, and it would give you the right thing.” It was prescient, as the vision has become a reality. But it is colliding with the business models that enabled Google, and publishers, to thrive.
Generative AI search appeals to most users because it feels human, fast and contextual. Instead of scanning headlines and toggling between pages, users get summaries. Google’s AI Overviews, OpenAI’s ChatGPT search plugins, and emerging players like Perplexity have tapped into this shift. Elizabeth Reid, Google’s head of Search, explained, “Human curiosity is boundless and people ask a lot of questions.” She added that search queries were becoming longer, conversational and multimodal, which links with the work that AI can do. Publishers argue that AI platforms are “free-riding” on their reporting (and content) without helping them get traffic. When the AI tool or search generates an answer, the incentive for the users to click the accompanying links vanishes. For the advertising-driven digital media houses, this implies fewer visits, and reduced eyeballs. It impacts the now-growing subscriptions, and shrinks advertising and affiliate revenues. Penske’s lawsuit is a response to this double-squeeze.
The financial stakes are massive. The advertising revenues from the search business of Alphabet, Google’s parent, are more than $175 billion a year. AI Overviews can help protect the market share against ChatGPT and Perplexity. The inherent risk is that the former may cannibalise Google’s existing search model. At the same time, generative AI startups have registered staggering valuations.
Perplexity, which is pitched as an “answer engine,” raised $63 million in early 2024, which pushed its valuation past $1 billion. OpenAI, thanks to its partnership with Microsoft, has reshaped searches, and productivity tools. Investors feel that “AI-first” search experiences will become the default for Gen Z, and subsequent generations.
Publishers find themselves on the debit side of the balance sheet. Advertising revenues that flowed steadily from the links-driven searches are drying up. While some like Penske have resorted to court cases to demand licensing deals from the AI-search tools, others like The Associated Press and Financial Times have struck agreements with AI companies to secure payments in exchange for access to their vast archives. But smaller outlets lack the clout to negotiate.
Ian Goodfellow, one of AI’s most influential researchers, said, “Generative models are changing the way we think about machine intelligence and creativity, and have the potential to transform industries from media to finance to healthcare.” This is creating new winners, which include cloud providers like Microsoft and Amazon, AI chipmakers like Nvidia, and startups with proprietary models. The losers include publishers, SEO-driven businesses, and smaller search players like DuckDuckGo, which find themselves a bit irrelevant.
No one knows the future of search in this scenario. A possible trajectory is the licensing economy, where AI firms pay to use publishers’ archives. Another is a hybrid model, where AI-generated summaries become the norm but are laced with richer links, nudges and citations to encourage users to click them. The third is that the publishers will adapt to develop “Generative Engine Optimisation (GEO).”
GEO means tailoring headlines, metadata, and even writing styles so that the content surfaces inside the AI summaries and answers. Like SEO (Search Engine Optimisation) created an entire industry in the early 2000s, GEO may be the next frontier. Many media houses feel that the future lies in subscription-heavy models. Despite AI, loyal readers will pay for trusted reporting, even as advertising revenues decline. Like viewers pay for OTT subscriptions, readers will not mind subscribing to favorite media outlets for news.
Regulators may step in to compensate publishers fairly. Else, if AI summaries falter with factual errors or oversimplifications, users may rediscover the value of Google search. Rather than a single destination or model, the journey seems more like reaching a crossroads with several forks, and diverging paths. The business disruption may mirror what happened during the transition from cable to streaming services in the mass entertainment sector. Incumbents scrambled, new giants emerged, and business models were rewritten.