As the misbegotten year that was 2023, which was a horrible year all around for my family, shambles reluctantly to a close, the better to usher in the New Year of 2024 (which, given the possibility of its bringing us Donald Trump as President again, threatens to be even worse than its predecessor), I was wondering what I could write about for one last post. I remember that I said that I would try to reinvigorate the blog next year, and I will. However, I’m unlikely to get to the task before January 2 or 3. In the meantime, though, I realized that I had totally forgotten to address a rather amazing tidbit that I encountered a couple of weeks ago on a quack website that has, alas, provided me with way too much material over the last two decades and has continued to descend ever further from just quackery and antivax misinformation into more generalized antiscience misinformation and conspiracy theories and just plain conspiracy theories, often mixed liberally with a hefty helping of prepper paranoia. I’m referring, of course, to Mike Adams and his misinformation empire Natural News, where three weeks ago I saw an announcement, Mike Adams announces breakthrough AI project that will bypass Big Tech censorship for health, nutrition and natural medicine:
In 120 days, or not later than the end of March 2024, an open-source, freely downloadable and locally run large language model (LLM) system that will use artificial intelligence tools will be launched. Said repository of natural health, nutrition, herbs, alternative medicine and modalities, disease prevention as well as mind, body and energy medicine information will be easily accessed in a user-friendly interface, without the need to go online. Thus, content is less likely manipulated, controlled and censored by Big Tech, Big Government and Big Pharma.
Brighteon and Natural News founder Mike Adams announced this game-changer news in a recent episode of the “Health Ranger Report.” “We have begun a project to build an AI chatbot, similar to chat GPT that is based on open-source systems that have matured substantially over the last two years,” he said. The data will be compiled from millions of documents that his team has acquired over the years, written by over 1000 authors. “And you will be able to get answers to questions right there as if you were talking to the expert authors of this content,” he told his podcast listeners.
My first thought was that millions of documents do not represent a very large dataset to develop an AI chatbot. The original ChatGPT model, for instance, was trained on an “immense dataset of internet-sourced data (570 gigabytes of text and 175 billion parameters.”
I didn’t watch all 50 minutes of his video announcing his “AI,” but if you are interested (and for completeness’ sake), here you go:
How is it that I didn’t comment on this when I first saw it? I don’t know. There’s just so much nonsense and misinformation and so little me these days. My first reaction was something along the lines of: Of course, Mike Adams is going to try to get into large language models (LLM) and artificial intelligence (AI). Remember his history. He’s long run his empire of websites and server farms, having become a master of black hat SEO in the process, but that’s not all. More importantly he got his start as a grifter running fear mongering Y2K scams 25 years ago and then graduated to forming a company selling email marketing software that helped clients doing mass email advertising to bypass the spam filters of their targeted recipients and did tens of millions of dollars a year in business. (Ah, those were so much more simpler times, weren’t they, when email spam was viewed as one of the worst threats to the Internet!) Then, when big tech social media platforms like Facebook started deplatforming and banning him, Adams pivoted to try to create his own alternative search engine and ecosystem online in 2019 that served up quackery and antivax misinformation, as well as his own Facebook-like social media platform.
So it’s a natural fit, given his past, that Mike Adams would leap onto the bandwagon of AI. Of course, any AI chatbot, regardless of the platform or LLM on which it’s based, depends a lot on what the input used to train it is, and the key passage above is: “The data will be compiled from millions of documents that his team has acquired over the years, written by over 1000 authors.” In other words, it sure looks to me as though Adams plans to train his AI chatbot on the contents of all his websites. I was also particularly amused by the message shown before the video even started:
Unsurprisingly, Adams starts out with a rant about how “big pharma” and “big tech” have captured the Internet and social media and how poor, poor pitiful quacks like him are being canceled. (Never mind that Adams runs an Internet and video empire with a massive audience to whom he can say whatever he wants.) Adams also rants how these companies also dominate search engines and the incipient efforts to develop AI that broke into the mainstream this year, thanks to Microsoft and OpenAI’s ChatGPT and Google’s Bard, among others, to which Apple will almost certainly add its contribution in 2024. He also hates Wikipedia, because of course he does.
While Adams has a point about big tech and big pharma’s dominance, his purpose is not to move towards a more democratic and science-based information ecosystem online, although he certainly cloaks his new project in this sort of language. Rather, as always, Adams’ purpose is to sell quackery and make money, and anything science-based is a threat. The source—big pharma, academia, the government, physicians, scientists—doesn’t matter. It’s a threat. In fact, even competing worldviews that might not be fully science-based represent a threat to Adams’ profit-driven misinformation and disinformation model, which might explain why in his introduction he makes a typically hyperbolic claim that there is an “ongoing effort to eradicate human knowledge about health and nutrition,” all to push people into vaccines, pharmaceuticals and psychiatric drugs. It’s basically more evidence of how, when it comes to cranks and conspiracy theorists like Adams, every accusation is a confession, because there is an effort to eradicate human knowledge about health and nutrition, at least science-based knowledge, to replace it with conspiracy theory-, pseudoscience-, and grift-based knowledge of human health and nutrition. Adams is at the forefront of that effort, along with people like Joe Mercola, Robert F. Kennedy, Jr., and others. Indeed, he proclaims that his goal in developing his AI chatbot is to be to make Google “utterly worthless” for searching for health and medical information online.
Amusingly, Adams admits that his 2019 attempt to develop a “natural health” search engine didn’t go that well, having apparently had some “challenges” with scaling. He does, however, try to claim why his effort will be “different” from Google without actually saying that it will be used on garbage-in to produce garbage-out:
He went on to point out that what he is building is far different from Google and other search engines, which are for the eradication of human knowledge. “A search engine doesn’t really answer that question. It only gives you a list of pages where you might go and perhaps you’ll find the answers there. But as you know, Google, Bing and even the other search engines wiped out the truth about health, nutrition and alternative medicine as they basically just push Big Pharma now,” he noted. The award-winning independent journalist also stated that search engines are designed to destroy knowledge and not to index it and make it available. “Google is controlled by the deep state, just as Wikipedia. Wikipedia is a tool of the CIA to control the public narratives about everything that matters and about all the people that matter,” he added highlighting that sites like these cannot be relied on to give access to information that can protect and enhance the quality of life.
I do love that part about Wikipedia being a tool of the CIA to control public narratives. I suppose I should be grateful that he didn’t say Wikipedia is a tool of the Illuminati or the Lizard People to achieve world domination; for Adams, that’s restraint. In any event, he claims that his AI chatbot will be able to accomplish this:
Also, he raised how his new offline platform is going to bypass censorship and other entities that could be violating one’s privacy. “Censorship is not the way forward. If you want to have prosperity, ingenuity, innovation, creativity, ideas, debates and democracy, it has to be based on the freedom to speak. And the ability to speak has to be powered by the ability to access knowledge,” Adams added. “The push for censorship right now is very loud and strong all over the world. You have to censor anybody who says anything critical about vaccines and you have to censor anybody calling to stop the bombing in Gaza. It’s just crazy. That just makes us less free,” he further emphasized.
To which I respond:
But what, exactly, does Adams plan to do? This, if you believe him (which is rarely a good idea):
Fifteen years ago, Adams created the website NutrientReference.com, because it has been his long-term dream to build a system that would provide all the information everyone needs to live a healthy life.
“I built this system back in 2007 or 2008, as what would now be described as a kind of a large language model. Basically, it was a language processing system that I personally wrote the code for,” he recalled, adding that the system was actually written inside Microsoft Visual Basic inside an enterprise SQL system, where he analyzed the text of over 1000 books that his team scanned. They used the text to begin parsing the language of different things out of the text.
“For example, I generated a list of what are all the known nutrients. And then there’s another list of what are all the known diseases and symptoms and another list of what are all the known organs or organ systems. And then I defined those and wrote this very complex system to go in and statistically, begin to link nutrients with diseases and so on,” he explained. He ended up publishing a system that analyzes all PubMed papers. However, the site did not gain much attention as people didn’t really find much value in it. According to him, it doesn’t work in the way that people like to ask questions.
“People don’t actually like to drill down through hierarchical datasets in order to find answers. Most people just want to ask questions in a common language,” he said. In the years since then, he has seen a revolution in AI systems, and more recently, ChatGPT.
And so he tapped the said “future technology,” only he’s going to make it based on open sources. “The bottom line is I want you to have all this knowledge at your fingertips. And you will, you will be able to install this and download it and run it on your own local computer,” the Health Ranger concluded.
My first thought reading this was to go and check out NutrientReference.com, which I encourage you to do. It’s actually not all that impressive in that it only appears to include a few hundred topics, nutrients, supplements, and medical conditions. Also, while 1,000 books scanned might have been impressive in 2007 or 2008, these days that’s a tiny number. Technology has advanced remarkably. I also found it amusing that Adams had to admit that his replacement for the PubMed database search engine (which I forgot to mention earlier) wasn’t all that popular and well-received.
I could go on and on and on about how Mike Adam’s ChatGPT-like AI chatbot is just another example of him glomming onto whatever is popular in technology and Internet search and social media and trying to create his own Mirror Universe version of it to sell his quackery, antivaccine ideology, conspiracy theories, and, of course, products ranging from supplements to prepper supplies, but I’ll wrap it up here. You can imagine for yourself what his AI chatbot will be like when it is released in March, assuming he hits his target date. What I wonder is whether Adams is just renaming his old search algorithms, rechristening them as “AI” even though they’re just old wine in a new AI bottle.
On the other hand, Adams could be part of a vanguard of ideological motivated cranks with tech skills who will go on to create AIs for specific purposes suited to their pseudoscience, conspiracy theories, and ideologies. After all, an AI is only as good as the dataset that it’s trained on, regardless of its algorithms. We already know that Adams’ dataset is small and certainly very biased; so we know what kind of results that it will likely produce. The questions that remain are how influential will this be, and will there be a bandwagon of special-purpose AIs designed to produce specific types of misinformation? Hell, unfortunately, we all know the answer to that last question. Truly, it’s going to be an even more horrible world when it comes to combatting pseudoscience, quackery, misinformation, and disinformation. Adams’ “AI” project is just one example of why.