Science Friday with Dr. Gustafson -- AI models and myth busting
10/16/2026
Welcome to Science Friday. Today’s topic is “Illusory Truth Effect,” where people believe something is true only because it is repeated so often, regardless of whether it is based in fact or supported by evidence.
???¼??™? ??™???™???™™??™???™???™? ??™–??™???™™ ??™???™®??™???™? ???½??™???™???™???™???™???™?
I was watching a University generated YouTube video on forage this week. It was five years old but still relevant, until the narrator said that there were two types of laminitis caused by grazing: One caused by hyperinsulinemia and another caused by eating forage fructans, and I quote, “… there’s another way that horses can develop pasture laminitis from carbohydrates that’s not related to hyperinsulinemia. This is related to cool season grasses and fructans,” followed by a lengthy explanation of endotoxemia induced laminitis from fructan overload.
One thing that frustrates me to no end is the lack of evidence to support this claim in the natural setting. To date, the only fructan that caused laminitis was inulin, derived from chicory root. A minimum of 5 lbs delivered in rapid bolus was required to induce laminitis in 100% of cases. The overload triggered endotoxemia. This is a dire clinical emergency with a clinical presentation of rapid heart rate, fever, lactic acidemia, a pathological drop in fecal pH, explosive diarrhea, ultimately culminating in laminitis (but no effect on insulin).
These experimental results were extrapolated to real life, with no evidence. Great leaps of faith are required – equating commercially available inulin with naturally occurring grass fructans, assuming an equine can consume that much fructan from pasture and translating the risk to hay. Never has such a clinical scenario been reported from eating high fructan pasture or hay, yet if you ask AI models, “Does fructan cause laminitis?”, you get back a most assured “Yes!” with all sorts of warnings about pasture and hay. I asked the AI model for evidence (none), then asked why it continuously gave an incorrect answer. Here’s the response:
“Your observation is scientifically accurate: there is no definitive experimental evidence proving that grass fructans cause laminitis or endotoxemia in horses under natural pasture conditions. AI models frequently claim otherwise because they are trained on a vast corpus of internet data. For over two decades, equestrian blogs, feed company marketing, general articles (and YouTube videos!) have oversimplified or misconstrued early veterinary studies, creating a massive digital echo chamber. When an AI processes this text, it reflects the sheer volume of this online repetition—essentially treating widespread veterinary folklore as a settled fact.”
Ouch.
I’ve generated two graphs, one for hay and one for pasture. This is ONLY about fructan, not HC (ESC + starch) that is typically in higher concentration than fructans. To calculate estimated fructan, subtract ESC from WSC.
The graphs show the percentage of fructan, how many pounds of forage would have to be consumed to take in the 5 lbs of estimated fructan theoretically necessary to trigger endotoxemia leading to laminitis. I added the realistic daily intake for an 1100 lb animal (green bar). I set the pasture moisture at 80% and the hay moisture at 8%. To reach 5 lbs of fructan intake, the forage would have to be between 20-35% fructan alone, the NSC would be much higher. Is this possible? Of course, anything is possible and there are samples that test that high, but is it probable? No. In my collective databases, no samples are this high. In another database specifically selected for NSC >20%, 7 of 780 samples had 20% or more estimated fructan alone. That’s less than 1% of “high NSC” hay.
So why doesn’t this “other way that horses can develop pasture laminitis from carbohydrates that’s not related to hyperinsulinemia” exist? There are 4 possibilities:
1. The experimental fructan and natural fructans are not equal
2. The chances of eating forage with >20% fructan alone are negligible
3. The horse cannot physically consume that much forage
4. It simply doesn’t exist
Next week we’ll investigate the ways fructans might be beneficial.
Kathleen Gustafson, PhD
President and Research Advisor, ECIR Group Inc
