You're asking AI for a competitive edge.
You're asking AI for a competitive edge. So is your competitor. You're both getting the exact same answer. Researchers at Esade, NYU Stern, and the University of Sydney just tested this. They ran 15,000+ simulations across 7 major AI mode…
You're asking AI for a competitive edge. So is your competitor. You're both getting the exact same answer.
Researchers at Esade, NYU Stern, and the University of Sydney just tested this. They ran 15,000+ simulations across 7 major AI models, including GPT-5, Claude, Gemini, and Grok.
The result? They coined a term for it: "Trendslop."
63% of the time, every model recommended the same trendy strategy: differentiation over cost leadership, collaboration over competition, long-term over short-term, regardless of the business context.
Only 12% of the time did a model recommend the less fashionable (but often correct) approach.
The kicker: "better prompting" barely helps. Adding rich industry context only shifted recommendations by 11%. The strongest biases moved less than 2%.
Why? These models are trained on internet text. "Innovation" and "differentiation" appear in positive contexts millions of times. "Cost leadership" and "centralization" appear in negative ones. The models aren't reasoning about your business, they're pattern-matching on cultural sentiment.
As the researchers put it: AI acts like "a freshly minted MBA, parroting what's popular rather than what's right."
Meanwhile, Walmart, Costco, Aldi, and Ryanair built empires on the strategy AI would never recommend.
This doesn't mean stop using AI for strategy. It means:
- Use it to generate options, not make decisions - Actively prompt for the unfashionable side ("Make the strongest case for cost leadership") - Treat "do both" recommendations as a red flag, not wisdom - Remember: if the advice sounds obvious, everyone else is hearing it too
The tool that was supposed to give you an edge might be slowly making you the same as everyone else.