From Links to Answers: A New Battle Over Credit
AI search citations describe how systems like ChatGPT, Gemini, Google AI Overviews, and Perplexity decide which sources to reference or hide when they answer user questions, shifting search from visible links to opaque attribution models that control who receives credit, trust, and traffic on the modern web. For nearly two decades, search meant typing a query and scrolling through ranked links. Now, people ask ChatGPT, Gemini, or Perplexity and receive a ready-made answer with little need to click away. In standalone AI chatbot traffic, ChatGPT holds 79.05% share, while Google still commands 90.39% of the wider search market, showing that one company owns the AI habit and the other owns the gateway. The collision of those roles, and how each tool credits or omits sources, is redefining how publishers, brands, and independent creators compete for attention.
ChatGPT vs Gemini vs Perplexity: Three AI Attribution Models
AI attribution models are already diverging sharply. In cooking searches, ChatGPT often returns a full recipe with no citation, as if the system invented the dish. Gemini AI Mode moves to the opposite extreme, typically citing between 16 and 29 sources for a single query and naming chefs, food bloggers, and specific sites it draws from. Perplexity occupies a middle ground, acting as an “answer engine” that names individual creators and links directly to their pages while still placing its synthesized answer first. In one test, Perplexity highlighted a specific recipe from Once Upon a Chef and listed ten sources, while Gemini cited 16 named sites and surfaced side panels of recipes from several well-known food blogs. These differences are invisible to most users but crucial for understanding who receives brand exposure, backlinks, and long-term audience growth.
AI Overviews and the Collapse of Click-Through
Google’s Gemini-powered AI Overviews show how search engine optimization AI is colliding with a new default: answers first, links second. Instead of “ten blue links,” AI Overviews pull information from multiple sites, blend it into a cohesive summary, and pin that text above traditional organic results. Studies cited in industry reporting show that when an AI Overview appears, click-through rates to underlying websites fall, because the overview satisfies most user intent on the page itself. ChatGPT and Gemini chats follow the same pattern: when a conversational response feels complete, there is little motivation to visit original articles. For publishers that spent years optimizing title tags, schema, and backlinks to win page-one rankings, the AI Overviews traffic impact is blunt. Ranking high still matters, but as an input to AI synthesis rather than a guarantee of direct visits.
Fragmented Search, Fragmented Traffic
The rise of AI assistants means the traditional search funnel has split into competing interfaces and citation strategies. Some users still start with classic Google results, but many now open ChatGPT first; others rely on Perplexity’s cited answers or Gemini integrated inside Google Search. Each tool fragments the web differently. ChatGPT absorbs content into uncited synthesis. Perplexity sends targeted traffic to a handful of named sources. Gemini sprays credit widely, calling out dozens of sites but reducing each to a small ingredient in its overview. This fragmentation is especially visible in recipes, where the “recipe internet” is splitting into ChatGPT-generated dishes, Perplexity-linked creators, and Gemini-summarized collections. For users, search feels smoother and more conversational. For the sites that funded that knowledge with years of publishing, visibility now depends on which AI front door people choose.
SEO in an AI-First World: Uncertainty for Publishers
As AI becomes the primary search interface, search engine optimization AI is shifting from ranking for keywords to becoming the preferred ingredient in AI answers. Content now competes for inclusion inside retrieval-augmented generation pipelines that select and summarize “relevant and trustworthy” pages. Yet creators rarely know when or how they are cited, if at all. In cooking, AI tools provide weeknight meal plans, adapt to pantry constraints, and, in Gemini’s case, even ask clarifying questions like “Do you have a specific protein on hand?” The experience feels closer to a personal assistant than a search engine. Meanwhile, many publishers report declining referral traffic when AI Overviews or chatbots satisfy user intent on their own. The long-term risk is clear: expertise may still power AI, but the visibility, loyalty, and clicks that once rewarded that expertise are far from guaranteed.
