A Cal Poly Pomona link redirects to a weight-loss scam using Oprah's name

By Zendoric · September 11, 2026.

The link circulating under the 'scams and deepfakes' tag appears to point to Cal Poly Pomona's virtual tour tool (its 'pannellum' panorama viewer). But a configuration parameter inside that same URL diverts the traffic to a different domain, bsoda.info, which hosts a page titled 'Baking Soda Shot, Oprah Baking Soda Recipe for Weight Loss? I Checked — Scam Breakdown'. We have not followed the redirect to its final destination: doing so adds no verifiable facts and would only reinforce the search-engine ranking such pages are chasing. What is an observable fact, and verifiable in the structure of the link itself, is the mechanism: an open redirect hosted on a university domain, used to carry traffic to a third party unconnected to that institution.

In general, the 'Oprah-endorsed baking soda weight-loss shot' is one more variant of one of the best-documented advertising frauds of the past decade in the United States: pages posing as press or independent analysis ('I checked', 'scam breakdown') in order to sell weight-loss supplements or gummies, leaning on the name of a celebrity who never gave consent. Oprah Winfrey has publicly denied ads of this kind on several occasions, and the US Federal Trade Commission (FTC) has pursued similar networks for exploiting her image without authorization. The 'scam breakdown' headline is no accident: it is the same ploy these networks use to disguise themselves as fact-checkers while, in reality, leading to the very product they claim to expose. Abusing trusted domains —universities, public administrations— to camouflage these links is a known technique for evading spam filters and search engines, one that exploits someone else's reputation without the affected institution bearing any responsibility for it.

Our take: this fraud pattern predates generative AI, but AI makes it cheaper in very concrete ways. Producing fake 'verification' pages, synthetic testimonials or, increasingly, generated voices and images imitating a real celebrity at volume costs today a fraction of what it did five years ago. It is exactly the kind of short-term cost we should not play down in our thesis: the same technology that promises abundance and better health also lowers the barrier to entry for anyone manufacturing fraud on an industrial scale, and it leaves collateral reputational damage at institutions —a university, for instance— whose domain is used without permission. The flip side is that automated detection of abusive redirects and fraudulent content is advancing along the same technological curve; today, however, the advantage still lies with whoever produces the deception, not with whoever detects it.

This piece is deliberately short: the available material is a link and a headline, not a verified article, and we prefer the honesty of a brief analysis to padding it out with details we cannot confirm.