# AI trust and the affiliate problem in generative search by John Morabito · April 19, 2026 · 8 min read **TL;DR** - Consumers trust AI recommendations more than search results. That trust is premature. - AI systems save compute by recycling top-ranking listicle articles. Those listicles are often pay-to-play. - Scaled "Top [Category]" AI-generated content is gaming the system. Nobody has heard of the "winners." - Google corrected similar SEO spam in September 2024. AI answers will eventually. Not yet. - Legitimate brands should invest in both strategic placements and genuine expertise content. ## AI systems take shortcuts These systems are engineered to conserve compute. When you prompt "best project management tool," the AI does not read hundreds of pages of reviews. It queries for existing listicles, clicks the top-ranking ones, and recycles the recommendations. The brands at the top of a Forbes or consumer-magazine roundup end up at the top of ChatGPT's answer. Not because the AI evaluated them. Because it read the same listicle you could have. ## The listicles are pay-to-play In the beauty and consumer space, major publications operate listicle placements as revenue. A brand buys $50K in advertising and gets a "Best of" placement as a value-add. The affiliate model incentivizes recommending products with the highest payouts, not the highest quality. The chain: brand pays for placement, listicle ranks well, AI reads listicle, consumer gets a "recommendation." Every link involves money. None of that context reaches the consumer. ## Manipulation is rampant and it works One company published hundreds of AI-generated "Top [Category]" articles, placing themselves at #1 across categories they had no expertise in. At industry conferences, audiences of seasoned professionals have never heard of these "winners." They still top AI answers. This works because Reciprocal Rank Fusion synthesizes across subqueries. Flood enough subqueries with self-promotional content and the AI treats you as consensus. ## The spam will get corrected. Eventually. Google has corrected similar patterns before. The September 2024 update crushed several scaled-content offenders on the SEO side. "Eventually" is doing a lot of work. There is a window. Brands playing honestly face a disadvantage in that window. Brands building on substantiated claims (real validation, real reviews, genuine expertise) survive algorithm corrections. Spam-builders are on borrowed time. ## What legitimate brands should do 1. **Influence the AI strategically.** No penalty for placements in cited sources. Digital PR, earned media, and strategic affiliate partnerships are legitimate GEO tactics. 2. **On-site content matters more than ever.** State your differentiators in crawlable text. Not PDFs, not images, not JavaScript widgets. 3. **Off-site corroboration is the differentiator.** Press coverage, YouTube reviews, social presence, industry citations. The corroboration layer decides who gets recommended. 4. **Invest in genuine-expertise content.** SME interviews, original research, case studies with real data. This is what survives algorithm corrections. ## FAQ **Are AI recommendations objective?** No. AI systems recycle top-ranking listicle articles. Those listicles are often pay-to-play. **Is there a penalty for gaming AI recommendations?** Not in AI answers currently. Google has penalized scaled content on the SEO side (September 2024). No equivalent enforcement yet in the AI layer. **What should legitimate brands do?** Both: strategic placements into cited sources AND genuine expertise content. Not either/or.