Pod: Why Should We Trust You?

John Wiley & Sons is at it again with its “date anyone with a cute AI” approach, announcing yesterday a deal with OpenEvidence (OE). This is nothing new for Wiley, a company that seems more infatuated with tech than it is with science and scholarship, a massive departure from its roots.

OE is certainly happy to make the deal, but not for the quality of the content. Rather, it keeps their targeted ad business growing. Platforming medical and health content is simply laying the groundwork for making a Facebook of medical information — a platform designed to sell user-targeted ads with the quality of content barely registering as a concern.

OE has proven susceptible to eugenics-adjacent pay-to-play misinformation, pay-to-play articles placed to help shill colostrum, and pay-to-play articles published to promote “functional medicine,” a MAHA-related pseudo-specialty. And nobody at OE seems to care, because that’s just part of the Section 230 platform game — don’t interfere in the content, just use it to get qualified leads to interact with your ad system.

While Facebook leveraged the trust of your social network — friends, family, neighbors, acquaintances — to get users into their ad platform, OE is leveraging trust markers in medical science, from brands to the concepts of peer review and evidence.

In addition to Cochrane, other trusted brands have been co-opted, including JAMA and NEJM, with house ads from both filling out pages on site:

Why these trusted epistemic brands are painting OE’s fence isn’t entirely clear, but it certainly smells commercial in nature, as if adoption of OE will in some way redound with positive financial returns to both NEJM and JAMA.

They aren’t the only ones using their trusted reputations to shill for OE, as we’ve noted before, calling out Bob Wachter for an unusually awkward set of promotions he swears amount to no financial interest. Yet, here he is, also via an on-site ad with his academic affiliation being used as another trust marker, along with the JAMA logo for some reason:

But there are numerous problems in the literature OE is promoting as reliable, including a recent set of 123 corrections for fictional case reports about infants and children which have led to a dangerous and troublesome vector of misinformation in the literature, one which OE reproduces without a qualm.

Can an LLM retract an article? Can an LLM correct an article, especially given how we can present correction notices? What happens when that article has permeated the scientific literature so that PubMed, ResearchGate, preprint servers, guidelines, and later research can be found citing or hosting it?

There is also the fact that OE isn’t stable enough to give you the same answer to the same query within minutes. This is not the first time something like this has been reported to me, and some of the answers diverge wildly from one another. Here are two examples of answers related to an earlier post, both of which are still wrong, but notice that they also don’t line up in the way a textbook entry might:

Bottom line? OE is an advertising platform being rolled out to target physicians with premium targeted ads, likely selling for a cost-per-view (CPV) of $500-$800. With major agencies backing it and others suffering FOMO, OE is positioned to clean up. However, with only about $7 billion in annual pharma ad spending, is this enough to justify their massive valuation? And will they be able to unseat actual point-of-care tools like UpToDate? Or will they just be a second-rate sideshow after the dust settles, the fundamental flaws become clear, and the short-term wins have been pocketed?

We also share our “Discoveries of the Week.”

• Fish doorbells: https://www.scientificamerican.com/article/how-the-dutch-fish-doorbell-helps-migrating-fish-each-spring/

Subscribe to our podcast

• Subscribe on Apple Podcasts
• Subscribe on Spotify
• Subscribe on Amazon Music/Audible
• Subscribe on YouTube

Music provided by Provoke the Truth — https://provokethetruth.net/