Introduction

Weeknight dinners often come down to whatever is in the fridge, and in my house that usually means chicken. Like many home cooks, I have leaned on AI tools to suggest ingredient swaps, rescue mistakes, and plan meals on short notice. So when The New York Times Cooking app introduced its own AI-powered assistant, I wanted to see if it could earn a permanent spot in my kitchen routine.

What Happened

The assistant lives inside the NYT Cooking app as a small text box labeled "Customize this recipe." It is currently available to subscribers on iOS. Instead of editing the original recipe, the tool generates a modified copy that highlights every change. I spent several days testing it on ingredient substitutions, recipe scaling, and equipment adjustments.

My first real-world test was a One-Pot Chicken and Rice With Ginger recipe. I lacked chicken thighs, chard, cilantro, and scallions, but I had tenders, green beans, red onions, and Thai basil. The assistant quickly rewrote the dish, adjusted the browning time, moved the green beans to steam over the rice, and approved the basil swap. My husband gave the finished meal his approval.

I also compared the tool with Alexa Plus on an Echo Show 11. Amazon's assistant offered similar substitution advice, but it could not display the modified recipe and eventually gave confused instructions. The NYT Cooking app proved more reliable, even if typing with messy fingers is less convenient than voice commands.

Things got trickier when I moved to baking. I asked the assistant to convert a Cinnamon Roll Bundt Cake recipe by editor Genevieve Ko into muffins. It adjusted the cook time and divided the batter, but when I asked for a butter measurement, it pulled from the modified muffin version without saying so. That would have left me with six tablespoons instead of the three-quarter cup required for the original Bundt cake—enough to ruin the texture. The app does not let you save custom versions separately, which adds to the confusion.

Why This Matters

Recipe purists have reason to be cautious. NYT Cooking is built on tested recipes from trusted sources, and some AI-generated recipes I have tried in the past were disappointing. This assistant is trained partly on reader comments and is programmed to reject requests that would produce bad-tasting food. Still, it is not infallible, especially with baking chemistry.

For everyday cooking problems—missing cilantro, the wrong cut of meat, or no Bundt pan—the tool can save a trip to the store and reduce guesswork. It also reflects a broader trend of AI entering recipe apps, raising questions about how much home cooks should trust automated guidance versus their own judgment.

Key Takeaways

  • The NYT Cooking AI assistant creates modified recipe copies rather than altering originals.
  • It handled weeknight chicken substitutions and ingredient swaps effectively.
  • Alexa Plus offered similar advice but could not show the updated recipe steps.
  • Baking conversions exposed a version-control flaw: the app gave muffin-recipe butter amounts without clarifying which version it referenced.
  • Custom recipes cannot be saved separately, so users should delete versions after cooking to avoid mix-ups.

Conclusion

The NYT Cooking assistant works best as a careful sous chef, not a head chef. It is genuinely useful for swapping ingredients and scaling dinner recipes, but bakers and anyone working with precise measurements should double-check every number against the original. If you are a recipe purist, you can simply scroll past the chat box. For me, standing in front of the fridge with a pack of chicken tenders and no plan, it has become a helpful part of the weeknight routine.