Overuse of “Misinformation” Label Complicates Scientific Debate, Expert Warns

When United States Health Secretary Robert F. Kennedy Jr. released new dietary guidelines earlier this year under the “Make America Healthy Again” initiative, the response from health organizations was notably divided. The American Heart Association praised the renewed focus on vegetables, fruits, and whole grains, while other groups expressed concern about recommendations promoting red meat and whole-fat dairy products. Some critics went further, accusing Kennedy of spreading “blatant misinformation” regarding the classification of butter and beef tallow as healthy fats.

This controversy highlights a growing problem in public discourse, according to Mu Zhu, associate dean of AI strategy in the faculty of mathematics at the University of Waterloo. The term “misinformation,” while critically important when addressing genuine falsehoods that threaten democracy, public health, and safety, is increasingly being misapplied to legitimate differences of scientific interpretation.

“There seems to be a growing tendency for people to apply the label to just about anything that they may disagree with, rather than genuine lies,” Zhu notes.

The professor of statistics points to the inherent difficulty in evaluating scientific evidence as a key factor in this trend. Nutritional science proves particularly challenging, as isolating the specific health effects of individual foods is complicated by countless genetic and lifestyle factors that influence outcomes. This complexity explains why many research studies can only establish “associations” or “correlations” rather than definitive causation.

Even in more straightforward cases, assessing evidence remains surprisingly difficult. Zhu offers a simple example: if a die shows an odd number six times out of seven rolls, does this constitute evidence of a loaded die? The answer depends entirely on which statistical approach is used to evaluate the data.

Using the traditional p-value method, which remains the most common statistical tool in science despite growing criticism, one might conclude there’s insufficient evidence to suggest the die is loaded. However, using an alternative approach called e-value, which examines relative likelihoods, one could reasonably conclude the die probably is loaded.

“The two arguments are not fundamentally at odds with each other but, using different thresholds, one ends up saying ‘black’ and the other ‘white’ when reality is just a certain shade of grey,” Zhu explains. While statisticians can calibrate these different methods to reach consistent conclusions, most people lack this technical ability when interpreting scientific findings.

This statistical nuance gets lost in everyday discussions about health risks. For instance, consumers might react very differently to being told that a food makes them “25 times more likely” to develop cancer versus learning it increases their cancer risk “from 0.01 percent to 0.25 percent” – despite these statements describing the identical risk.

Neither interpretation is inherently wrong, but the current climate has made expressing certain viewpoints increasingly risky. “Today, I’m afraid to say: ‘I see no need to change my diet given the risks,'” Zhu admits. “If I did, those who are enthusiastic about changing theirs might come together and accuse me of spreading ‘misinformation.'”

This trend threatens productive scientific discourse. Zhu argues that the term “misinformation” should be reserved exclusively for genuine falsehoods, not for conclusions that differ based on subjective statistical thresholds, even when those thresholds represent conventional choices. Relying solely on traditional statistical significance thresholds has already generated numerous irreproducible scientific findings.

The stakes are particularly high as concerns about AI potentially worsening misinformation grow. Without more nuanced approaches to evaluating evidence and claims, we risk further polarizing discussions around complex scientific topics like nutrition, health, and environmental issues.

“To call something ‘misinformation’ based on that kind of shaky evidence, or the lack thereof, will only obstruct true scientific progress,” Zhu concludes. As dietary debates like those surrounding Kennedy’s guidelines continue, more precise language and greater tolerance for legitimate scientific disagreement may prove essential for productive public health discussions.