Stanford researcher Ellen Kuhl estimates that there are approximately 10^43 potential burger recipes worldwide. With BurgerAI, a new tool developed in her lab, artificial intelligence can now design a burger tailored to an individual's age, taste preferences, nutritional needs, and sustainability goals.
BurgerAI's capabilities extend beyond simply suggesting appealing or nutritious burgers. According to Kuhl, the innovation represents a broader shift for artificial intelligence from prediction to design. "Most AI systems are trained to predict what already exists. We wanted AI to invent what should exist next," said Kuhl, who is a professor of mechanical engineering in the School of Engineering and directs Stanford Bio-X. "BurgerAI does not ask, 'What burger is most likely?' It asks, 'What burger best satisfies these important and complex objectives?'"
Kuhl said food was chosen as the focus because it combines elements of human experience and culture with health and environmental impact—topics that attract multidisciplinary researchers across various Stanford schools. Vahidullah Tac, a Schmidt Science postdoctoral fellow in Kuhl's lab, said, "Food choices are some of the most consequential decisions humans make every day... With one arrow, you can hit two targets – planetary health and personal health. It's a great and impactful research area." The team has published two papers on BurgerAI; one introduces the tool while another connects its mathematical principles with those underpinning diffusion-based generative AI used in fields such as materials design and physics.
To develop BurgerAI’s recommendations, the team used 2,216 burger recipes from Food.com as data for learning ingredient combinations and quantities before generating new recipes from scratch. These novel recipes were then matched against human flavor profiles for preference optimization based on gender, age, and physical activity levels.
The system underwent real-world testing when five professionally prepared AI-designed burgers were served in a blinded taste test at a San Francisco restaurant involving more than 100 diners. In comparisons with popular fast-food burgers, BurgerAI’s creations scored equally or better on overall liking as well as flavor and texture metrics. Notably, its Mushroom Burger demonstrated over tenfold reduction in environmental impact compared to standard options while its Bean Burger achieved about twice the nutritional score.
Kuhl said, "AI did not just generate plausible burger recipes – it created burgers that real people enjoy... That may sound simple but it means the model learned what makes food appealing to the human palate." Tac added, "We expected some trade-off between sustainability and consumer acceptance... But we found a burger with dramatically lower environmental impact could still compete with one of the world's most successful burgers." Both researchers view BurgerAI primarily as proof-of-concept for generative AI’s potential role in designing solutions across fields such as pharmaceuticals or advanced materials where balancing multiple objectives is crucial.
