WASHINGTON, D.C. — The New York Times’ Cooking app launched a beta AI‑driven “Recipe Modifier” this week, letting users swap out ingredients, adjust servings and accommodate dietary restrictions with a single tap. Early testing shows the tool generally delivers accurate substitutions while preserving flavor profiles.

The feature taps a proprietary large‑language model trained on the Times’ archive of 20,000 recipes. Users enter a change—such as replacing dairy with oat milk—and the AI rewrites the method, recalculates nutritional information and suggests timing tweaks. In internal trials, 84 % of the 12,000 beta participants said the output matched their expectations.

KEY POINTThe feature taps a proprietary large‑language model trained on the Times’ archive of 20,000 recipes.

Enterprise interest is already emerging. Food‑service operators like Compass Group have begun pilots to integrate the modifier into employee‑wellness portals, hoping to personalize meals for 1.2 million staff worldwide. “Scalable, on‑the‑fly recipe adaptation could reduce food waste by up to 15 % in corporate cafeterias,” said Maya Patel, senior director of product innovation at Compass.

Competitors are moving fast. Allrecipes recently unveiled a similar tool powered by OpenAI’s GPT‑4, while startup Whisk offers a subscription API for grocery retailers. Analysts note that the NYT’s advantage lies in its editorial authority, but “the market will reward the most seamless integrations,” warned tech analyst Jorge Alvarez of Forrester.

“The AI respects the culinary intent of the original author, which is a big step beyond generic ingredient swaps,” observed chef‑author Thomas Keller, who contributed recipes to the Times. User data shows a 22 % increase in repeat sessions among those who tried the modifier, suggesting strong engagement potential.

Key Takeaways

• The NYT Cooking AI modifier is in beta, with over 12,000 active testers.
• Enterprise pilots aim to cut corporate food‑waste by up to 15 %.
• Competitive pressure is rising as other media and tech firms launch similar AI tools.

While the feature is still polishing edge cases—such as complex cultural spices—the early response indicates a viable path for AI to personalize home cooking at scale, and a new frontier for media brands to monetize editorial expertise.