Overview
- UCL researchers, in a study published Friday in Scientific Reports, unveiled a validated system that scores diet and nutrition posts by the risk they pose.
- Called Diet-MisRAT, the rule-based model adapts the World Health Organization’s hazardous-exposure methods to online content and assigns green, amber, or red risk tiers.
- The team calibrated the tool across five verification rounds against judgments from nearly 60 specialists in dietetics, nutrition, and public health, finding it delivered reliable results.
- The framework flags inaccuracy, hazardous omissions, and manipulative framing as key traits of high-risk content and classifies claims pushing high-dose vitamin A over the MMR vaccine as critical risk.
- The authors say risk scoring can guide safeguards for platforms, policymakers, and AI developers as harmful advice persists online, though coverage has not reported formal adoption.