Part of the reason for forming SBM in the first place is because my colleagues and I perceived that we were engaged partly in a long struggle for control over information – specifically the quality of health information being pushed to the public, including the media and regulators. We did not anticipate how much this battle would change over the last few decades, and we honestly struggle to keep up, but overall I think we’re fighting the good fight.
Initially the campaign to promote science-based medicine was largely framed as opposing the incursion of “alternative medicine” into the mainstream. This fight mainly took place within academia, and so one of our early instruments was a print medical journal, The Scientific Review of Alternative Medicine, run by Wally Sampson. The journal itself was excellent, but we eventually realized this wasn’t where the fight was actually happening. It was more at the public opinion and regulatory level. Further, the people who really needed to read the journal, mainstream health experts not already enamored of CAM, were not interested.
The ultimate death of SRAM is what gave birth to SBM. This was close to the onset of web 2.0 and social media, and I sensed this was where the struggle would focus, and I think I was correct. An online format was more immediate, more relevant, and could reach more people. Of course, the snake oil peddlers sensed this too, and proliferated.
It then became critical to have SBM articles rank highly in search engine results on any relevant topic. Here the outcome was mixed. We did well (and continue to do well) because Google and other search engines realized they need to weight results based on quality to produce a good search product. This meant that links from more reliable sources counted more. No one knows for sure exactly what algorithms they use, but much of the time an SBM or skeptical article would rank highly on most searches of medical topics. But – we simply were increasingly outnumbered.
The next line for the information battle was Wikipedia, an increasingly dominant online reference source for all information. Wikipedia is actually a clear success for science-based information of all kinds, and we have the editors to thank for that. They found a way to contain the endless editing wars, and clearly value science-based and rational information over pseudoscience and self-promotion. Some groups tried to start competitors, but they amounted to little more than vanity projects. Wikipedia rules, and is definitely in the win column for science and SBM.
But the information ecosystem continued to evolve. Social media has increasingly dominated, and it also has proliferated, with new platforms each developing their own style and subculture. It is not enough to have a blog that is an excellent resource for science-based medicine topics. We have to expand to all corners of social media. We do that to some extent, but this is an endless project and there is much to be done.
I have taken up some of the slack with my personal TikTok feed, which is mostly SBM topics. There is a culture there of conspiracy mongering, anti-expertise, and the closest venue I am aware of that can be said to be “post-truth”. My weekly whack-a-mole over there is better than nothing, but I feel like it’s a drop in the bucket. We need an infusion of young talented SBM promoters as force multipliers, and to take up the banner of SBM once the current writers eventually age out.
So, where is the next battle line for the promotion of SBM over pseudoscience and snake oil? I think it’s AI. What may end-up mattering the most in a short time is what answer popular chat-bots give to medical questions, especially ones relevant to SBM. I am keeping a close eye on this, and it’s a mixed bag as you might expect. A lot of the information that chat-bots spit out comes from Wikipedia, which is a good thing. But also their output depends a lot on how the prompt is phrased. Here too volume matters – how much science vs nonsense is in the training data.
This does mean that all that SBM content we have been creating over the last two decades is likely included in the pretraining data of most models. But then again – so is all the CAM promotional pseudoscience and propaganda.
I just prompted Chat-GPT with “what do you think about homeopathy?” It responded:
“My assessment of homeopathy is that it is a pseudoscientific system of medicine whose central principles are inconsistent with established chemistry, physics, and physiology, and whose clinical efficacy has not been demonstrated beyond placebo effects.”
That’s a pretty good response, but I chose homeopathy because it is low-hanging fruit. It gives a more mushy answer when asked about acupuncture, not a hard-hitting SBM answer. Still, overall not bad.
There is also the potential to make our own SBM chat-bot. You can take what’s called an unweighted pretrained model, basically the first step in training a large language model, and then you can complete the training on curated information and give it guardrails to influence its behavior. We could take an open model and train it on all of SBM’s articles – essentially make an SBM chat-bot. This might be a good resource for people specifically looking for SBM information or doing research. I doubt, however, that it will gain widespread use. Like Wikipedia, we need to influence the main products out there, not just create our own specialized product. Still, worth exploring.
As always, we can benefit from volunteers willing to put some effort in to promote science in medicine, and fight the endless incursion of quacks and con-artists, who have been depressingly successful over the last four decades. The information wars will never end.
SBM And The Battle For Information