Water Fast Coaching and Articles

A disturbing trend

Over the last year, I’ve noticed a disturbing trend. An increasing number of clients are coming to me having self-diagnosed – and unfortunately often misdiagnosed – their health issues through internet ‘research’. (I’ll give some examples a little later.) In itself, this problem is hardly a new phenomenon. It’s even been acknowledged in the Netflix series The Pitt, with Dr. Robby confronted in one episode with parents who are convinced, after consultation with ‘Dr. Google’, that they know what’s better for their son than the head of the ER himself. I won’t reveal any spoilers, but needlessly to say it leads to complications…

If the situation is bad enough in standard medicine, then the dangers of using Dr. Google and Dr. ChatGPT are far greater when it comes to holistic approaches to health – and especially when it comes to AI’s recommendations for therapeutic fasting protocols. If that weren’t already bad enough, the trend has been getting a lot worse even within the last year.

From Google to ChatGPT

The first, and perhaps biggest, danger has been in the transition from self-diagnosis using a simple Google search to an unquestioning reliance on the wisdom of AI. In the ‘old days’ before AI, you ultimately had to take responsibility for making any self-diagnosis, decision or fasting protocol. You could get on the internet, type in a search term, and Google would spit out a million search results. Ultimately, however, it was your job to decide which articles to read, which articles to prioritise, and, for better or worse, reach a conclusion based on the content of what you read. Things have become very different in the last year. Now AI decides which articles to choose, which articles to prioritise, and – based on secret algorithms that will never be revealed to anyone – come to a conclusion about what is best and true for your health. All you have to do is sit back, stop thinking, and let AI do all the work…

Why do we assume that AI is infallible? It certainly isn’t.

Changing algorithms

The flawed assumption that ChatGPT must know everything is predicated on the fact that AI can only ever be as ‘smart’ as the resources it is allowed to learn from. Unfortunately, when it comes to the realm of fasting (or alternative medicine in general), there are severe limitations in its knowledge base. Back in 2024 Google began the trend towards lowering the rankings of alternative medicine sites, ostensibly in order to reduce the presence of ‘quacks’ on the internet. This also coincided with a more general trend, in which it began to noticeably favour search results from large corporations over small businesses. That trend has continued with the rise of AI, which is trained on mainstream medical thought, using mainstream research and articles as its knowledge base. Research articles which don’t appear on NIH/PubMed (effectively, the US government’s national library of medicine) are either downplayed or ignored altogether.

NIH/PubMed is a fantastic resource which I use myself on a weekly basis, but it doesn’t log journals or research articles which diverge from mainstream allopathic medical thought. For instance, naturopathic, ayurvedic, homeopathic journals are all but excluded, which means that AI doesn’t get to learn from them or benefit from respected traditions which go back hundreds or sometimes thousands of years.

Where does that leave AI’s knowledge about fasting? Or general attitude towards fasting? You’d might as well just ask your family doctor – and we all know what the average MD thinks about fasting!

The major limitations of mainstream scientific fasting research

It’s true that today there are plenty of research articles about fasting in mainstream medical journals, and AI does pick up on this. However, there are multiple problems:

• Most mainstream research articles are in vitro (test tube) studies and not in vivo (in an actual live body)
• Most of the in vivo studies which do exist are based on animal studies – primarily mice – and not human studies.
• Most mainstream research articles examine only one or two specific biochemical markers, ignoring the full ‘recipe’ of hundreds if not thousands of peptides, enzymes, neurotransmitters, hormones etc. which all interact with one another in a living body, whether human or animal.
• Most mainstream research articles only take ‘snapshots’ at predetermined time intervals, rather than observing the fluid and evolving biochemical environment which exists in a live human body.
• Perhaps most fundamentally, most mainstream articles base their research on relatively short periods of fasting: usually up to only 3-5 days in the case of in vivo human studies, given that anything longer must surely be dangerous…

This all means that the information available to AI isn’t even applicable to the longer therapeutic fasts in which healing can develop more significantly. This blinkered, short-sighted approach is what allows influencers to come up with pseudo-scientific soundbites such as Mindy Peltz’s ‘autophagy begins at 17 hours’. Not 16 hours. Not 18 hours. Should we really take this number as gospel, when a lab technician simply made an observation at a randomly assigned point in time? More fundamentally, however, this pseudo-precise timeframe also ignores the fact that the gradually evolving process of autophagy inevitably varies from person to person: something I observe clinically with every client.

AI: Fasting ‘makes cancer worse’

Let’s get more specific, with one of the many examples of AI confusion in the world of therapeutic fasting.

I’ve had a few clients come to me with the worrying assertion, based on their AI research, that fasting makes cancer worse. I decided to check this online myself by typing in the search phrase: ‘Does autophagy make cancer worse.’ Google’s AI overview came back with the following result:

Yes, autophagy can make cancer worse. While it prevents cancer in healthy cells by cleaning up damaged material, established tumors hijack autophagy to survive stress, resist treatments, and spread.

Yikes.

If you accept AI at face value, it looks like fasting must be dangerous. Better not do it! And unfortunately for the average person, that’s as far as you can go in trying to make sense of it all. However, the fatal flaws in AI’s conclusions become immediately obvious with a little specialist understanding, along with a little digging beneath the surface. The two articles that Google AI quoted were:

(1) https://www.pratikpatil.co.in/is-autophagy-good-or-bad-for-cancer/ :
A general webpage by a mainstream doctor with ‘12+ years of experience in treating solid and blood cancers’ using mainstream drug-based interventions.

(2) https://pmc.ncbi.nlm.nih.gov/articles/PMC5066235/ :
‘Recent insights into the function of autophagy in cancer’: a research article from the journal Genes & Development. According to the abstract, the article studies ‘how autophagy can promote cancer through suppressing p53 and preventing energy crisis, cell death, senescence, and an anti-tumor immune response. Keywords: autophagy, ATG, cancer, mouse models, chloroquine.’

To an average person, the first article is accessible, but the second one reads like gobbledy-gook. With greater specialist knowledge, however, it turns out that both articles are inapplicable to the clinical reality of extended therapeutic fasting. The problem originates in the term ‘autophagy’ itself and how we interpret its meaning. To a lab researcher, autophagy refers specifically and exclusively to intracellular processes (achieved through organelles such as lysosomes) which help the cell break down and remove internal toxins and waste products. This contrasts significantly with the use of the word in fasting parlance, which refers more broadly to any kind of ‘auto-phagy’: that is, ‘self-eating’, as per the word’s etymology. From this point of view, intracellular autophagy is just one element of the whole package. From a fasting perspective, therefore, autophagy also includes upregulated immune system cells (such as macrophages and natural killer cells) which maraud through the body during fasting, seeking out and destroying entire cells which lack the biochemical ‘ID’ to prove that they are healthy and happy. Of course, it is this latter form of autophagy which can enable a fasting body to heal from and ‘autophagise’ scar tissue, fibroids, cysts and, yes, tumors.

In this particular case, AI jumped to dubious conclusions based on the misinterpretation and misapplication of a single keyword. What other misunderstandings are possible next time you ask ChatGPT a question?

More fundamental concerns about the use of AI for self-diagnosis…

AI also serves up more general, flawed information about fasting protocols (including the role of electrolytes), dietary preparation and refeeding. Don’t get me started! Even more worrying than this, however, is the trend I’ve seen among clients who falsely self-diagnose with AI and then self-prescribe inappropriate medications or supplements. This can and often does lead to a further deterioration of their health – which is why they’ve signed up for a consultation with me in the first place. Understandably, the average person is going to ‘research’ their individual symptoms at face value, failing to make underlying connections between those individual symptoms which reveal the true nature of their illness. The bottom line is that if you feed the wrong information into AI, it can only possibly give you the wrong answer. In other words, not only is AI fallible in itself, but it also has no chance to provide accurate information when provided with humanly fallible search questions and phrases.

Just in the last week I’ve had consultations with two clients who misdiagnosed or mistreated their health issues based on their internet research with AI. One client had caused a flare-up of her gut issues through the misapplication of herbal anti-microbials, while the other had prescribed herself a hormonal cream which, over the long term, was only going to deepen her hormonal imbalance.

Where is this all going?

I honestly don’t understand why people take such huge risks with their health. Why do we want to believe that AI knows everything? Have we really benefited by throwing out the omniscient gods of old, in favour of a new god based on the One, True Science? Partly, of course, it’s in the interest of tech companies to hype AI’s potential. Fair enough, that’s all part of the marketing game. Above and beyond this, though, there’s a certain comfort for many people in handing over control of their health to someone else. In the old days it was a human doctor. ‘Just tell me what to do Doc!’ Now it’s Dr. ChatGPT.

I deeply believe that taking responsibility for our own health and illness is the first step on the road to recovery – as well as to a fuller life in general. But getting expert advice along the way is crucial! Fasting is a perfect example of this. It doesn’t just happen to you. You take responsibility, you take a deep breath. And then, having dotted all your ‘I’s and crossed all your ‘T’s, you go ahead and do it.

(This article was written and edited the old-fashioned way: without the help of ChatGPT.)