The University of Illinois Urbana-Champaign reveals how AI, specifically ChatGPT, is used in the sensory evaluation of brownies, potentially transforming food product development.
Artificial intelligence is becoming an integral part of various industries, revolutionizing the way we approach numerous tasks. But could AI change how we evaluate the taste and quality of food, particularly something as universally cherished as brownies? A new study from the University of Illinois Urbana-Champaign, published in the journal Foods, explored just that.
The research by Damir Torrico, an assistant professor in the Department of Food Science and Human Nutrition of the College of Agricultural, Consumer and Environmental Sciences at Illinois, delves into how ChatGPT can be harnessed for the sensory evaluation of foods. Specifically, Torrico’s study focused on brownies, evaluating whether AI can effectively mimic human responses to culinary delights.
The process of sensory evaluation is vital in the food industry. Before new products hit the market, companies rely on trained human testers and large consumer panels to assess flavor, texture and overall appeal. This procedure, however, is not only time-consuming and costly but also prone to human biases and fatigue.
“Sometimes, relying on human testers can slow down the process, especially when multiple product prototypes need to be evaluated simultaneously. Sensory panels require time and careful coordination, and in some cases, certain ingredients may not be food-grade, making them unsuitable for tasting,” Torrico said in a news release. “That is why large language models such as ChatGPT are being considered for sensory evaluation. It is possible to create models that can replicate certain human responses.”
In his study, Torrico provided ChatGPT with 15 different brownie recipes, including conventional and unconventional ingredients, such as mealworm powder and fish oil. The AI was tasked with describing the sensory characteristics of each brownie in terms of taste, texture and overall enjoyment, with responses subsequently categorized as positive, negative or neutral.
Remarkably, ChatGPT’s evaluations were overwhelmingly positive, irrespective of the ingredients. This phenomenon aligns with the concept of hedonic asymmetry, where both humans and AI tend to describe beneficial experiences in positive terms.
“ChatGPT was trying to always see the good side of things,” Torrico added.
Despite this optimistic bias, Torrico believes that AI has a promising future in food development. ChatGPT could serve as an initial screening tool, helping food scientists narrow down recipe options before human evaluation, significantly saving time and resources.
“Using AI can give general insights of what products can be considered for further testing, and what products shouldn’t be put through that long process,” added Torrico. “I could see ChatGPT being developed for sensory evaluation to help the industry.”
While AI might not replace your favorite brownie-tasting session just yet, Torrico is optimistic about its potential. Future plans include refining the experiment to train ChatGPT to use vocabulary that closely matches that of human sensory panels.