On November 8, 2026, the Hebrew University of Jerusalem announced that a machine‑learning framework had sifted through 50.7 million chemical molecules to identify 92 promising sugar‑substitute candidates, according to a statement released by the university’s research group. The announcement, reported by Xinhua, highlighted that the AI models not only predicted sweet and bitter taste profiles, but also simulated how the candidate molecules might bind to the single human sweet‑taste receptor, a critical step toward developing low‑calorie sweeteners.
The research team began by compiling three extensive chemical databases, collectively containing around fifty‑seven million distinct compounds. Using known sweet‑ and bitter‑tasting molecules as reference points, they trained predictive models to differentiate desirable sweetness from undesirable bitterness or metallic aftertastes. This approach allowed the system to screen the entire set of compounds in a fraction of the time traditionally required for laboratory testing.
Initial screening reduced the pool to 556 potential sweeteners. The researchers subsequently applied advanced molecular docking simulations, taking advantage of a detailed structure of the target sweet‑taste receptor that had become available last year. This step assessed how each candidate would engage with the receptor’s binding pocket, a key factor influencing perceived sweetness and potential side‑effects. After refining the results based on binding energy and predicted taste outcomes, the team settled on 92 molecules that exhibited favorable sweet‑taste receptor interactions while avoiding predicted bitter profiles.
Importantly, several of the selected compounds feature chemical structures distinct from existing sweeteners currently used in food and beverage products. This novelty suggests the possibility of entirely new classes of sugar alternatives that could offer reduced caloric content without the aftertaste drawbacks that have plagued many alternatives over the past decades.
The study, published in the journal Food Chemistry, underscores how artificial intelligence and molecular modeling can accelerate the discovery of food additives. By automating the early phases of screening and evaluation, researchers can focus laboratory efforts on a far smaller, more promising set of candidates, potentially shortening the timeline from concept to market approval.
From a public health perspective, the availability of new low‑calorie sweeteners could influence consumer choices related to sugar consumption—a major concern for many governments and health organizations worldwide. The reduction of bitter or metallic aftertastes is particularly relevant, as unpleasant sensory experiences often lead consumers to abandon alternative sweetener products in favor of regular sugar, thereby undermining long‑term dietary benefits.
Industry stakeholders have expressed long‑standing interest in developing sweeteners that combine high sweetness potency with neutral taste profiles. The Hebrew University’s announcement marks a potentially significant advance in meeting that demand, given the efficiency and scale of the AI approach.
While the study did not detail regulatory pathways, the formulation of a rigorous computational pipeline establishes a foundation that could streamline future compliance reviews. By demonstrating the predictive power of AI in taste profiling and receptor binding, the research signals a shift toward data‑driven food science that could impact product development cycles nationwide.
The findings also raise questions about intellectual property and access to AI‑generated compounds. Although the announcement was limited to describing the scientific approach and candidate list, the potential for patents on novel sweetener structures remains a factor that could shape commercial strategies in the food industry.
Overall, the Hebrew University’s AI‑driven discovery of 92 new sugar‑alternative candidates represents a notable milestone in applying machine learning to food chemistry. By addressing a persistent sensory challenge—bitter aftertastes—the research opens a path toward healthier, more palatable sweetener options that could benefit public health and consumer choice alike.
