What the new AI-based research is revealing about rice, starch, and glycemic response
- Whole-grain rice (brown, red, purple) is not always low glycemic—grain color doesn’t reliably predict blood sugar impact, which can vary widely across varieties.
- Starch structure inside the grain (chain length, crystallinity, and interactions with proteins/lipids) is the key factor determining how quickly rice raises glucose levels.
- AI and machine learning are transforming rice research, enabling faster, cheaper identification of low-GI varieties and helping breeders develop healthier rice for managing rising diabetes rates.
By Rhowell Tiozon Jr., Myrtel Valenzuela
For many years, brown, red, and purple rice have been widely viewed as healthier choices compared with white rice, especially for people concerned about blood sugar. The logic has been simple. Whole grains contain more fiber and nutrients, so they should also lead to slower digestion and lower glycemic impact.
However, emerging research suggests that the picture is more nuanced.
As diabetes continues to rise globally, affecting hundreds of millions of people today and projected to increase further in the coming decades, there is growing interest in understanding how staple foods like rice influence blood sugar in more precise ways. Yet even among health-conscious consumers, experiences often vary. Some whole-grain rice still appears to trigger noticeable glucose spikes.
A recent study published in Carbohydrate Polymers and led by researchers from the International Rice Research Institute (IRRI) and De La Salle University (DLSU) offers new insight into this question. Using machine learning tools, researchers examined 384 whole grain rice varieties to understand what truly determines how quickly rice raises blood sugar. The findings show that glycemic response is governed by multiple structural and compositional features of the grain, such as amylose content, amylopectin chain length distribution, starch crystallinity, resistant starch formation, and interactions between starch, proteins, and lipids. The present study shows that specific starch chain-length patterns are particularly predictive of glycemic response across diverse rice varieties.
Color is not a reliable indicator of blood sugar response
One of the clearest results from the study is that grain color, whether brown, red, purple, or white, is not a consistent predictor of the glycemic index (GI).
Across the large set of varieties analyzed, GI values ranged from approximately 60 (intermediate) to over 90 (high) across the evaluated rice varieties, with substantial overlap between pigmented and non-pigmented groups. Some pigmented varieties showed relatively high-GI values, while some non-pigmented types performed better than expected.
This does not reduce the value of colored rice in terms of fiber, antioxidants, and other nutrients. However, the study found that these antioxidant properties do not appear to strongly influence how quickly starch is digested and converted into glucose.
The researchers noted that differences in blood sugar response are more closely linked to the starch structure rather than external grain characteristics such as pigmentation.
The role of starch structure inside the grain
Rice starch is made up of long chains of glucose units arranged in highly organized structures. The study found that the length and arrangement of these chains play a key role in determining how quickly enzymes can break starch down during digestion.
In simple terms, some starch structures are more tightly packed and harder to break apart, leading to a slower release of glucose into the bloodstream.
The researchers identified certain chain lengths that were associated with lower glycemic response, particularly short-length chains (DP 33-36) and specific medium-length chains that help form a more compact and stable starch structure. These structures appear to slow down enzymatic digestion.
Other chain patterns were associated with faster digestion and higher glycemic response.
Rather than a single good or bad trait, it is the combination and arrangement of these starch components that shape how rice behaves nutritionally.
Why rice type and structure also matter
The study also highlighted differences between major rice groups, such as Japonica, often short-grain and sticky, and Indica, typically long-grain varieties such as Basmati-type rice.
In some cases, polishing, which removes the bran layer to produce white rice, resulted in higher glycemic responses, particularly among many Japonica varieties. However, the effect was more variable in Indica genetic backgrounds. In several lines, glycemic responses remained relatively stable after polishing, and in a few cases even showed slight improvements. This suggests that the internal starch structure of the grain can sometimes play a stronger role than the presence or absence of the bran layer alone. While brown rice typically retains more nutrients because the bran layer is preserved, some brown rice varieties can still produce relatively high glycemic responses depending on their starch architecture.
How AI helps researchers find low-GI rice
Traditionally, identifying rice varieties with a low glycemic index relied on human trials, which are slow and expensive. IRRI researchers found a faster approach using artificial neural networks. These AI tools analyze starch structural patterns and predict blood sugar responses without the need for immediate human testing. By combining detailed starch measurements with machine learning, researchers reduced screening costs from about US$1,000 per sample to less than US$10 per sample — a reduction of more than 99%.
With AI-assisted prediction, thousands of rice samples can be screened in advance, effectively doubling the speed of the discovery process and allowing only the most promising candidates to move forward to clinical testing. This technology enables breeders to develop rice varieties that better support metabolic health, an increasingly important goal as diabetes and diet-related diseases rise worldwide.
What this means for your plate
Whole-grain rice offers important nutritional benefits, but it is not a guaranteed shield against blood sugar spikes. The hidden starch architecture inside each grain ultimately shapes how your body processes energy.
As rice science and AI continue to reveal the inner workings of rice grains, the humble bowl of rice is emerging as a precision tool for human health. Next time you serve yourself, the question should not be “Is it brown?” but “How is it built?”
Read the paper:
Putlih Adzra Pautong, Rhowell Navarro Tiozon Jr., Reuben James Q. Buenafe, Ana Rose Ramos-Castrosanto, Hyeyeon Kim, Gacel Perfinian, Glenn Vincent P. Ong, Il-Ryong Choi, Rodolfo Sumayao Jr., Nese Sreenivasulu
Starch fine structure predicts glycemic index variation in whole-grain rice
Carbohydrate Polymers, Volume 382, 2026
https://doi.org/10.1016/j.carbpol.2026.125236
