A team of American scientists has developed an AI-designed burger that aims to optimize taste, health benefits, and environmental sustainability. Using a database of more than 2,200 burger recipes, the researchers created a mushroom-based patty designed to have a significantly lower environmental impact than conventional fast food options.

The project was led by Professor Ellen Kuhl at Stanford University, where her lab developed the BurgerAI system. Unlike typical AI models that predict existing trends, BurgerAI generates new recipes by prioritizing multiple complex criteria, including flavor, nutrition, and ecological considerations. Prof. Kuhl emphasized that the tool seeks to identify what kind of burger "should exist next," rather than simply producing the most likely recipe.

The resulting mushroom patty combines portobello mushrooms, arugula (also known as rocket), rosemary, grains, and various condiments. In taste tests conducted at a San Francisco restaurant, over 100 participants sampled five different AI-generated burger variations. Two versions of the “Delicious Burger” created with BurgerAI scored comparable or higher than a popular fast-food burger in overall liking, flavor, and texture.

Environmentally, the mushroom-based burger reduced its impact by more than an order of magnitude compared to the fast food alternative, while a bean-based variant developed by the team achieved approximately twice the nutritional score. According to Prof. Kuhl, these outcomes demonstrate that AI can go beyond merely generating plausible recipes—it can create foods that consumers enjoy and that also contribute positively to health and sustainability goals.

The research also highlights the potential for personalizing burger recipes to suit different taste preferences, leveraging AI’s capacity to balance diverse factors. Prof. Kuhl noted that food represents an emerging frontier in biosciences because it integrates aspects of culture, nutrition, and environmental impact, making it a complex, multifaceted challenge that AI methods are well suited to address.