Sometimes, the answer from an AI to a question that you pose is problematic. The reality may have much more nuance than the answer you get, or you could even have a full-fledged moment of “AI hallucination” on your hands. And today, I saw a prime example of how some truth and some overstatement was mixed with outright false information, to produce an answer that misled the AI user and stirred up a lot of confusion.

Here is the information that I saw posted on a Facebook group:

I immediately recognized the things that had been grouped together, and which couldn’t be supported. But the (well-intentioned) user couldn’t see what was wrong. This is where AI can get us in trouble, because answers are provided with a high degree of what reads to the human as “confidence”, but the actual veracity may be sorely lacking.

Let’s look at the issues in this answer, and lay them out clearly.

1. “Magnesium chloride absorbs more easily through the skin”

Unfortunately, there is no convincing head-to-head human study showing that magnesium chloride is absorbed better through intact skin than magnesium sulfate. In fact, there is a lot of conflicting and contradictory information concerning topical magnesium use in general:

• A 2016 laboratory study using excised human skin found that some magnesium could penetrate the stratum corneum, particularly through hair follicles. (This is not the same as being absorbed generally by the skin.) That demonstrated the possibility of penetration, but not how much would reach circulation or tissues in a living person. 
• A small 2017 placebo-controlled cream study included only 25 people. The overall changes in serum and urinary magnesium were not statistically significant; significance appeared only after the investigators excluded four participants classified as athletes. One investigator also disclosed a potential future financial interest in a magnesium cream, which sorely impacts how unbiased the results might be.  
• A 2024 uncontrolled magnesium-chloride study involved only six people with ileostomies and chronic hypomagnesemia. Two had clear serum increases and one avoided a planned infusion, but this unusual population and tiny sample cannot establish ordinary skin absorption. We have too few study participants and a clearly specialized segment of people.
• Most importantly, a study published just days ago in June 2026 found minimal magnesium-sulfate permeation through intact porcine skin over 24 hours. Of course, pig skin is not human skin and so we still can’t extrapolate. But that said, permeation of the magnesium increased dramatically only after the skin barrier was experimentally stripped. The authors concluded that simple aqueous topical magnesium delivery is unlikely through normal intact skin. 

Having said all that, I do think that you can get magnesium support via the skin (because there have been thousands reporting benefits via the Trying Low Oxalates group (which can be found on Facebook and Groups.io) but “magnesium chloride absorbs more easily” is much more definitive than any of the evidence supports.

There may also be differences in concentration and formulation. Note that magnesium chloride hexahydrate (the substance in your typical “Magnesium oil” solution) contains slightly more elemental magnesium by weight than magnesium sulfate heptahydrate (which is the chemical name for Epsom salt). Application by sprays of magnesium oil may be much more concentrated than any epsom salt bath! That could also make these two products appear different without proving that chloride crosses skin more readily.

2. “Magnesium binds oxalates in the gut”

This is the strongest statement from the AI answer —but it applies to oral magnesium taken with food, not magnesium applied on the skin or absorbed from a bath. So this is actually adding a confusing aspect to the response, because the initial information had to do with two topical approaches.

That being said, there are benefits to taking oral magnesium, and binding oxalate in the gut is one of them! In a controlled human oxalate-loading study (where the test subjects were deliberately taking in a higher oxalate source), oxalate absorption was reduced by taking either a magnesium supplement or a calcium supplement:

• 13.5% oxalate absorption with the oxalate load alone
• 7.6% when magnesium oxide was consumed with the oxalate load
• 5.1% when calcium carbonate was consumed with the oxalate load.

As you can see, the oxalate absorption was reduced almost as much with the magnesium supplement (in this case magnesium oxide) as it was with the calcium supplement (calcium carbonate). However, this has nothing to do with topical magnesium!

Topical magnesium cannot produce this gut-binding effect. Why? We are not getting a concentrated dose to the gut. Even if some magnesium enters circulation through the skin, it has missed the intestinal compartment where dietary oxalate binding occurs. So the comments regarding magnesium’s benefits here (while valid) apply only to oral supplements.

The wording that magnesium “stops oxalates from crystallizing and forming kidney stones” is also too absolute. It may help in general, but it cannot provide perfect protection. Magnesium may:

1. Bind dietary oxalate in the intestine (assuming that the magnesium is available and “bumps into” the available oxalate).
2. Potentially bind with oxalate in the bloodstream, which may reduce the likelihood of calcium-oxalate crystallization in urine (but won’t necessarily prevent all stone formation).

Both types of magnesium binding with oxalate are useful mechanisms, but magnesium supplementation by itself has not consistently prevented recurrent stones in clinical trials. 

3. “Magnesium supports liver enzymes that process salicylates”

This is a very generic argument that magnesium is a cofactor for processing salicylate. But there is no specific salicylate mechanism which uses magnesium as a substrate. What do I mean by this?

First of all, magnesium is not a specific to salicylate clearance. In fact, salicylic acid is cleared mainly through:

• Glycine conjugation to salicyluric acid
• Glucuronidation
• Smaller oxidative pathways
• Renal excretion, which is strongly influenced by urinary pH

So your liver has work to do in terms of salicylate, as does your kidneys. Those pathways in each of these organs have been characterized in human pharmacokinetic studies. 

Where does magnesium come into this? Magnesium is involved in ATP-dependent metabolism throughout the body. If you have tissues that are trying to generate energy (which is what ATP is to the cell), then those tissues need magnesium. As a result, severe magnesium deficiency could impair many processes, but only indirectly. This is a far cry from demonstrating that added magnesium—particularly topical magnesium—could meaningfully accelerate salicylate clearance. At this point in time, there is no human research evidence that supports this idea.

4. “Magnesium reduces reactions to salicylates”

At this point, there are no controlled research evidence that magnesium treats dietary salicylate intolerance, aspirin-exacerbated respiratory disease, salicylate-triggered urticaria, or an MCAS-like salicylate phenotype. When looking at the research, we really have nothing at this point that could support this assertion!

It is quite possible that an individual may feel calmer or less reactive after magnesium because magnesium can influence neuromuscular excitability, sleep, stress perception and vascular tone. These are documented benefits that those who need magnesium may display when taking a magnesium supplement. Correcting low magnesium could also improve general physiological resilience, because we do need magnesium to produce ATP. But this would not be due to increasing salicylate metabolism or correcting the underlying reaction mechanism.

5. The “sulphur sensitivity” claim

There are a few issues here. But the biggest one is that sulphate is not interchangeable with sulphite (also spelled sulfite), sulphur-containing foods, sulphonamide medication, thiols or hydrogen sulphide. These are all very different substances, which are used and processed very differently in the body.

Magnesium sulphate contains sulphate, a stable, highly oxidized inorganic sulphur compound. A sulphite-sensitive person does not automatically react to sulphate, and sulphonamide allergy does not imply sulphate intolerance. Allergy specialists specifically caution against treating these chemically distinct categories as cross-reactive. 

That does not mean people won’t feel worse after an Epsom-salt bath. They may! But potential explanations could include:

• Bath temperature or duration
• Vasodilation or blood-pressure effects
• An unexpectedly large physiological shift
• Skin irritation
• The concentration of the bath
• A reaction to another product used simultaneously
• An individual response to sulphate that has not yet been well characterized

So, can we automatically assume a sulphur intolerance of some kind? No. Someone who consistently reacts to magnesium sulfate may reasonably try magnesium chloride, but “sulfur sensitivity” is not an adequate scientific explanation.

Magnesium chloride may also sting or irritate the skin more than a dilute Epsom-salt bath, especially on shaved, inflamed or damaged skin.

Where do we go from here?

Let’s put the whole thing in a better context, shall we? I’d sum up here with the following:

Magnesium can support general mineral status and nervous-system regulation. In fact, I generally recommend supplementing magnesium in some form, because of how oxalate acts as a mineral chelator. We are more likely to be depleted in our minerals, and if not dealing with outright deficiency, we may have at least some level of insufficiency.

Oral magnesium taken with meals may also reduce intestinal absorption of dietary oxalate. While we do know that anecdotally topical magnesium products seem to provide good support, evidence for clinically meaningful absorption of magnesium through intact skin is limited and conflicting. Further, while topical magnesium products seem to provide support to the body overall, there is no good evidence that topical magnesium binds oxalate, enhances salicylate clearance or specifically reduces salicylate sensitivity. Magnesium chloride and magnesium sulphate may be tried according to individual tolerance (and both should be started slowly at 1/4 to 1/2 cup in a bath, and increased to as much as 2 cups as long as no negatives have been seen), but neither has been proved superior over the other for absorption or specific conditions.

But there’s another aspect here – which is the use of AI in your search for good quality health information.

I personally do make use of AI but I do so very carefully – and I have purchased a subscription to more than one AI, and I will do a number of things to ensure the best quality information:

1. In any AI interaction, I’ll ask it to provide me with all the research links that the answer is based on (so I can crosscheck the information myself).
2. My instructions to the AI will always include instructions on whether it can hypothesize or not.
3. In every case, I ask the AI to provide me with a confidence level that the answer is accurate.
4. It’s important to fine tune every inquiry to put “guardrails” on any question, so that there is little room for the AI to head off in all directions!

But perhaps the most important thing in my use of AI is that I personally have a lot of context and knowledge in my head already – so I can spot more readily when an AI has given me a low quality or even false answer. So here’s what I’d say if you are using AI to get support in your health journey, and you aren’t getting better – it would be well worth consulting a human who might just have both heart and wisdom, that the AI may not have!

For more information from me on oxalate, minerals, supplements and all things about nutrient dense, low toxin eating, check out my YouTube channel.