The standardized measurement of AI-generated misinformation about real brands has, until this week, not existed as a publicly published framework in U.S. communications research. The 5W Hallucination Index, launched today with a fashion-category pilot, is the first formal attempt to operationalize the discipline.
The framework measures three dimensions of brand-risk error inside AI answers: accuracy (verifiable false statements about a brand's products, leadership, or history), drift (gradual reframing of a brand's positioning across engines over time), and brand hallucination risk (the rate at which an engine confidently invents details that did not happen). The pilot covers fashion. Subsequent indexes will cover pharmaceuticals, financial services, automotive, hospitality, and food and beverage — categories where hallucinations carry direct legal, medical, or financial consequences for consumers and brands.
The need isn't in dispute. AI research labs, regulatory bodies, and brand-safety vendors have flagged the scale of generative misinformation as a structural problem. What's been missing is a measurement layer specifically built for brand-side risk — one that produces a repeatable, category-comparable score communications leaders can present to their boards, general counsels, and risk committees.
"For twenty years the worst thing a brand could read about itself was on the front page of the Wall Street Journal," Ronn Torossian, founder and chairman of 5W, said in a statement coinciding with the launch. "Now it's inside an AI answer that fifty million people read in a week, and nobody at the brand even knows it's there. The Hallucination Index is the smoke detector."
Methodological questions worth watching: the comparability of hallucination rates across engines with different retrieval systems and confidence calibrations; whether brand-side measurement can scale to the speed at which models update; and the harder question of remediation — once a hallucination is identified, what is the playbook for surfacing the correction inside the engine? 5W's methodology indicates remediation runs through structured authority (Wikipedia, structured data, earned media) rather than direct engagement with the engines themselves, consistent with the practical state of the discipline in mid-2026 but likely to evolve as model providers build more direct correction pathways.
The Index also raises a procurement question for general counsels and chief risk officers: where does responsibility sit when an engine generates a false statement that causes consumer harm? Legal frameworks are still developing. The FTC, the European Commission, and several state attorneys general have signaled interest in AI-generated commercial misinformation. Brands that have measurement systems in place when regulatory frameworks tighten will be in materially better defensive position than brands that do not.
For the communications industry, the launch suggests a maturation of the AI visibility discipline beyond pure citation-share measurement. The first generation focused on inclusion: is the brand named in the answer? The second generation, which the Hallucination Index represents, focuses on accuracy: is what the engine says about the brand actually true? Both layers will need to be measured. Communications teams that build only the first will have a blind spot in the second.
The 5W Hallucination Index Fashion Pilot is available at 5wpr.com/ai-visibility-index/hallucination-index-fashion-pilot/.
