New York Times adds more AI features to its Cooking app. Users aren't thrilled
New York Times Cooking is adding generative AI features to suggest ingredient substitutions and adjust recipes, prompting sharp pushback from many paying subscribers. Critics argue that they pay specifically for rigorously tested, human-curated recipes and fear a slide into unvetted “AI slop,” as well as being forced to subsidize technology they dislike or distrust. Supporters counter that the tools are optional, can mirror workflows users already perform with standalone AI assistants, and—if executed carefully—could offer genuinely useful, context-aware guidance in the kitchen.
Overall reaction to NYT Cooking AI features
- Thread splits between users who find the feature promising and those who see it as “AI slop” invading a paid, human-curated product.
- Many critics emphasize they pay for tested, reliable recipes and don’t want untested AI variations mixed into that experience, even if optional.
Perceived usefulness of AI in cooking
- Supporters say ingredient substitutions, quantity scaling, and technique adjustments are good AI use cases, matching workflows they already do with general-purpose chatbots.
- Some frequent home cooks say they would use this regularly, especially for on-the-fly substitutions or adapting to available ingredients and equipment.
- Others argue they can already do this with their own AI tools, tuned to their tastes and constraints, so site-specific AI adds little.
Trust, quality, and brand concerns
- Strong worry that this is the first step toward cutting back expensive human test kitchens and relying on AI guesses.
- People emphasize the value proposition of NYT Cooking: tested recipes, curation, and cultural context, not auto-generated variations.
- AI hallucination/confabulation is cited as a serious risk, including unsafe instructions or spurious ingredients.
Consent, control, and UI design
- Critics want the ability to fully disable AI features, framing it as a moral and product-consent issue.
- Others respond that optional features don’t need per-user toggles: “just don’t click it.”
- There is concern about “AI everywhere” UX patterns (persistent buttons, attention-grabbing entry points) and comparisons to past intrusive features like Clippy or pinned Copilot buttons.
- Some suggest AI integrations should be subtle, clearly labeled as fallible, and focused on discovery or guidance rather than replacing core content.
Broader sentiment about AI and “slop”
- Several comments see this as part of wider “enshittification”: pushing AI where older, simpler solutions would work better.
- Counterpoint: low-quality, repetitive cooking content long predates AI; AI is just a new way to generate or refine it.
- Debate over terminology (“hallucination” vs. “confabulation”) reflects broader tension between AI enthusiasts and skeptics.