AI's Next Gold Rush: Influencing the Answers Everyone Sees

The author opens with a thought experiment about how reputation management has worked for the past twenty years. Companies hired SEO consultants, PR agencies, and crisis-management firms to manipulate search rankings—pushing negative stories off the first page of Google. The key point: Google never answered the question itself. It ranked documents. Users could always dig deeper, compare sources, and form their own conclusions. Google optimized discovery, not consensus.

Large language models changed that contract silently. An LLM doesn't rank documents—it synthesizes them. It reads thousands of articles, court filings, blog posts, Reddit threads, academic papers, and news reports, then compresses them into a single answer. For most users, that answer becomes the product. The sources become footnotes. Modern AI search systems like Perplexity or SearchGPT do expose citations, but the relationship has inverted: the links are no longer the primary product. They are supporting evidence for the synthesized answer. Most users do not audit the footnotes any more than they visited the second page of Google.

This seemingly small UX change creates an entirely different economy. SEO was built around influencing which documents people discovered. But what replaces SEO when the documents disappear? The author poses a scenario: imagine you're the CEO of a public company with safety issues and regulatory scrutiny. Millions of customers no longer start with Google—they ask ChatGPT, Claude, or whatever comes next. Would you rather spend a million dollars trying to convince dozens of journalists to rewrite their stories, or making sure the AI assistant everyone trusts produces a slightly different summary—not false, just more balanced, more contextual, more charitable, more 'aligned'? Nobody would market this as censorship. The product names would be 'Brand Safety,' 'Hallucination Reduction,' 'Context Optimization,' 'Responsible Responses.' The invoice would say 'We improved answer quality,' not 'We removed inconvenient facts.'

Zooming out, this incentive structure applies to any domain where AI becomes the default research tool: politicians, countries, medical treatments, investments, universities. If one synthesized answer becomes the interface to human knowledge, influencing that answer becomes one of the most valuable services in the world. The author acknowledges this sounds paranoid but notes that markets form around valuable bottlenecks—and a single synthesized answer is an extraordinarily valuable one. The harmless version of this industry is already emerging: SEO is turning into 'GEO' and 'LLM Optimization' as companies learn to structure content so AI systems preferentially consume and quote their material. The more interesting question is what happens once influencing public AI systems is worth billions instead of thousands. The author cites Anthropic's recent essay on open-weight models as evidence the debate is live, noting that economic incentives don't require bad intentions. The conclusion: as society converges around a small number of trusted AI systems, the most valuable product may not be intelligence itself, but the ability to influence the first—and often the only—answer millions of people ever see.

For two decades, reputation management operated at the margins of discovery. Companies hired consultants to manipulate search rankings, but Google's core product remained transparent: ranked links that users could compare and evaluate independently. The system had friction—and that friction was the point. Anyone motivated enough could verify sources, check page two, or cross-reference information.

Large language models have removed that friction entirely. An AI assistant doesn't show you ten competing sources and let you choose. It reads thousands of documents, newspapers, filings, and posts, then distills them into a single coherent answer. For the vast majority of users, that answer IS the product. The sources—if visible at all—are decoration.

That small technical shift has massive economic consequences. In Google's world, influencing information meant persuading publishers and journalists to bury or reframe stories. In an AI world, it means persuading (or paying) AI systems to synthesize information in ways subtly favorable to your interests. The language matters: nobody will sell this as censorship or propaganda. They will sell it as 'improved answer quality' and 'responsible deployment.' The economics are clean: if a single synthesized answer becomes the trusted default for millions of people asking about politics, products, medicine, or finance, controlling what that answer says becomes extraordinarily valuable. The article notes that this incentive structure—not malice—is what drives the market. Economic bottlenecks get monetized whether or not anyone intended them to be.

No comments yet. Be the first to share your thoughts!