AI Rewrites Your Recipes, But NYT Cooking Users Aren't Buying It

NYT Cooking, the cooking app from The New York Times, is currently testing an AI feature that allows users to modify existing recipes. Its supported capabilities include recommending ingredient substitutions, suggesting alternative cookware, and scaling recipes to any desired serving size. The official description positions the feature as a tool to make recipes "adapt to the cook's needs in that exact moment".

One demo example shows that a user wants to replace kale with baby spinach. The AI reminds the user that baby spinach wilts much faster, so it should be added in a much later cooking step.

The feature was announced on October 23 by Genevieve Ko, senior editor at NYT Cooking. According to The Verge's report, the rollout of the function has sparked dramatic reactions among users. The report did not list every specific point of contention, but the inherent tension behind the debate is obvious just by looking at the feature itself.

NYT Cooking is a subscription product. One of the core reasons users pay for access is their trust in the professional judgment of its test kitchen and editorial team. Recipes are not ordinary reading content: a 10-degree difference in cooking temperature, or a 5-minute gap in cooking time, can make the difference between an edible dish and a ruined mess. When the AI gives suggestions in this scenario, users are completely justified in asking: Has this suggestion been properly tested? How does the model know my oven performs exactly the same way as the editor's test oven?

Previously, operations like scaling serving sizes and swapping ingredients were adjustments that happened between editors and users. Now that an AI model is inserted into the middle, the chain of trust users have in the source of the cooking advice becomes much longer.

This does not mean the feature itself has no value. Swapping ingredients and adjusting serving sizes are indeed very frequent, high-demand needs. But NYT Cooking's core selling point is "reliable, trustworthy cooking advice", and users have very different expectations for this platform than they do for a generic public chatbot.

Nearly all content platforms are rushing to embed AI features into their products right now. NYT Cooking's move is hardly a radical step. The discussion it has sparked makes one thing clear: users do not necessarily oppose the presence of AI in products, they care deeply about what role the AI is allowed to take on. In users' minds, the human recipe developer and the AI model that tweaks recipes are two entirely distinct roles.