Texture Independently Drives Liking in AI-Generated Alternative Protein Burgers

Vahidullah Tac, Aeneas O. Koosis, Ellen Kuhl · 2026

Read Texture Independently Drives Liking in AI-Generated Alternative Protein Burgers by Vahidullah Tac, Aeneas O. Koosis, Ellen Kuhl (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-08365.txt.

Introduction

Vahidullah Tac, Aeneas Koosis, and Ellen Kuhl show that texture alone can drive how much people like a burger, using patties designed with artificial intelligence. The team paired a blind tasting with one hundred one participants against mechanical texture profile measurements for six burgers.

The lineup mixed AI-generated alternative protein patties with animal, plant, mushroom, and hybrid patties and a classic Big Mac. Animal burgers held a distinct sensory-mechanical zone tied to firm, fatty, and holds-together impressions. Mushroom patties skewed springy and gummy, while plant patties read as dry, brittle, and crumbly, in protein-dependent ways. The findings give product designers measurable texture targets linked to liking.

Catalog record

CREATORVahidullah Tac, Aeneas O. Koosis, Ellen Kuhl
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE34 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.08365

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Texture Independently Drives Liking in AI-Generated Alternative Protein Burgers Vahidullah Tac, Aeneas O. Koosis, Ellen Kuhl Abstract Texture shapes how we perceive and like food, yet clear links between mechanical measurements and sensory perception of texture remain elusive. Here we combine sensory data from a blind tasting with 101 participants with mechanical texture profile analysis across six burgers to identify the textural features that drive consumer perception and liking. We compare five burgers–generated with artificial intelligence–with animal-based, plant-based, mushroom-based, and hybrid animal-mushroom patties, and the classical Big Mac. Three main findings emerge: First, animal-based burgers occupy a distinctive and coherent sensory-mechanical region associated with attributes such as firm, fatty, and holds together. Second, mushroom- and plant-based burgers deviate from this region in protein-dependent ways: mushroom-based burgers associate with springy and gummy textures, while plant-based burgers associate with dry, brittle, and crumbly textures. Hybrid animal-mushroom burgers, however, maintain sensory profiles comparable to fully animal-based burgers. Third, resilience emerges as the strongest mechanical correlate of perceived meatiness and sensory texture, while stiffness and hardness show no statistically significant association with consumer perception. Texture independently predicts overall liking alongside flavor: increasing texture liking by one point increases overall liking by 0.28. Among all sensory attributes, meatiness is the dominant predictor of texture liking. These findings identify resilience as a promising target for texture engineering and establish texture as a critical design objective for sustainable alternative proteins. Correspondence: [email protected] Keywords: generative artificial intelligence; texture profile analysis; sensory survey; texture; resilience; meatiness 1 Motivation Few foods have shaped human health and environmental sustainability as strongly as burgers. Burgers concentrate many of the defining challenges of modern food systems into a single product: nutrition, sustainability, flavor, texture, and consumer perception [ 13 ] . Conventional meat production places a major burden on both human health and the environment [ 7 ] . Animal agriculture contributes substantially to greenhouse gas emissions, land use, and water consumption [ 18 ] . Diets rich in processed red meat also increase the risk of cardiovascular disease, metabolic disorders, and colorectal cancer [ 34 ] . Plant- and mushroom-based proteins offer a promising path toward healthier and more sustainable foods [ 15 ] . Yet consumer adoption remains limited by persistent gaps in flavor and texture relative to animal meat [ 23 ] . At the same time, generative artificial intelligence is emerging as a powerful tool for food design, capable of exploring large formulation spaces and optimizing recipes for nutrition, sustainability, and taste [ 27 ] . These advances create new opportunities to engineer alternative proteins with targeted sensory and mechanical properties [ 28 ] . Texture plays a central role in how we perceive and enjoy food [ 26 ] . In burgers, texture shapes sensory impressions such as softness, firmness, juiciness, fattiness, fibrousness, and meatiness, all of which strongly influence liking and purchase intent [ 32 ] . Texture exists on two fundamentally different levels [ 23 ] : Consumers perceive sensory texture through chewing and oral processing, while laboratory instruments measure mechanical texture through controlled deformation experiments [ 6 ] . Establishing quantitative links between these two domains remains challenging, particularly for alternative proteins [ 9 ] . Animal meat derives its characteristic texture from the hierarchical architecture of muscle fibers, connective tissue, and fat networks [ 19 ] ; plant- and mushroom-based meats instead form structurally different matrices with distinct deformation and fracture mechanisms [ 8 ] . As a result, many plant- and mushroom-based meats fail to reproduce the cohesive, elastic, and meat-like texture of animal meat [ 33 ] . This makes alternative-protein texture engineering a heterogeneous soft-materials problem in which microstructure, phase connectivity, loading rate, and coupled deformation mechanisms govern the effective mechanical response [ 3 , 31 ] . Quantifying these mechanical responses is therefore essential for understanding why different burger formulations produce distinct sensory experiences. Texture profile analysis provides a quantitative framework to characterize the mechanical texture of food through standardized and repeatable laboratory measurements [ 5 ] . The method applies a double-compression test that mimics the process of chewing and extracts parameters such as stiffness, hardness, cohesiveness, springiness, resilience, and chewiness from the resulting force-displacement response [ 4 ] . Unlike subjective sensory evaluations, texture profile analysis generates reproducible measurements under controlled loading amplitudes and rates and enables direct comparison across formulations [ 17 ] . However, the mechanical parameters from a texture profile analysis do not directly map onto human sensory perception, and it remains unclear which parameters best predict perceived texture and overall liking [ 9 ] . Here, we bridge this gap by integrating sensory surveys and texture profile analysis across six AI-designed burgers to characterize how mechanical texture shapes sensory perception and liking of animal-, plant-, mushroom-, and hybrid animal-mushroom burgers. Figure 1: Overview of the six burgers used in this study. We generated five new burgers using our generative artificial intelligence platform for burgers: delicious burger 1 and delicious burger 2 were optimized for palatability and contain animal meat; sustainable burger 1 was optimized for sustainability and uses mushroom; sustainable burger 2 uses a hybrid animal-mushroom blend; and nutritious burger uses beans. We use the classical Big Mac ® for benchmarking, it uses animal meat. 2 Materials and Methods We previously designed a generative artificial intelligence platform for food design [ 27 ] . We used this tool to generate five burgers and optimize for palatability in the animal-based delicious burgers 1 and 2, for environmental impact in the mushroom-based sustainable burger 1 and the hybrid animal-mushroom sustainable burger 2, and for nutritional value in the plant-based nutritious burger (Figure 1 ). We conducted a sensory survey with n= 101 respondents from the general population. Participants evaluated the textural properties, flavor, and overall liking of the five AI burgers and, as a benchmark comparison, the classical Big Mac ® . Here, we use the survey data to assess the effect of patty type on the perceived texture, investigate which textural attributes contribute most heavily to the participant’s liking of a burger, and explore how animal- and plant-based burgers compare along those attributes. We also perform texture profile analysis tests [ 17 , 24 , 33 ] of the six patties to quantify the texture of the burgers in the laboratory. We obtain texture profile analysis parameters such as stiffness, hardness and cohesiveness for each of the six burgers and investigate whether these mechanical attributes can be used to detect perceived textural attributes. 2.1 Generative AI for burgers Machine learning models and generative AI are increasingly used to learn compact, lower-dimensional representations of complex systems and design spaces; ranging from food safety to physical systems like materials modeling and sediment transport [ 20 , 2 , 16 , 30 ] . We use generative AI to learn human preferences in food, specifically burgers, and use the resulting model for targeted recipe generation for goals such as environmental sustainability and nutritional value. We represent each burger recipe in terms of the ingredient selection and ingredient quantification. Thus, burger recipes are hybrid discrete-continuous objects and modeling such data requires a two-stage approach that respects the hybrid nature of the data. We create a generative AI model that

Read the complete text