In short
- AI-AVATARS and influencers push a high dose of magnesium with misleading health claims.
- Doctors warn of real health risks, especially of vulnerable groups.
- Experts insist on consulting professionals and eating healthy food instead of taking more pills.
Magnesium has a moment. Recommended as the latest remedy alles by wellness influencers and their increasingly realistic AI-cloning-the mineral floods social media feeds with promises of better sleep, less stress and longer lives. But medical experts say that the trend is full of wrong information and is potentially dangerous.
Promoted by people and Avatars generated by Avatars about Tiktok, Instagram and YouTube, magnesium supplements are sold as catch-all healed for anxiety, insomnia and weight gain. Some supplements pack up to 700 mg per dose – almost twice the recommended maximum – and that has doctors who generate red flags.
“Ai” Doctors “give testimonials that are directly, black and white:” If you have this problem, this is how it works. ” It is convincing if you do not realize that it is AI, “said Dr. Steven Chen, associated dean for clinical affairs at the USC School of Pharmacy, said Decrypt. “What worries me about these rapid fragments is that they do not mention the risks.”
Chen said that although magnesium has legitimate use-such as managing the regularity of the intestine and the relief of muscle tension that benefits are picked and chosen exaggerated.
As AI-generated images become more advanced, experts say that the line between facts and fiction is more difficult to recognize.
Most adults do not need more than 400 mg of magnesium per day, and that is easily treated by foods such as peanut butter, cashew nuts, chia seeds, chicken fillet and salmon. But viral supplements are sold as a performance-fast, easy and approved algorithm, with some supplements that sell pills that offer more than twice that level.
Although Magnesium itself is not illegal or regulated as a medicine, the way it is sold by algorithm-driven promotion and generative AI witnesses has converted it into a profitable health product that is brought to the market with little supervision.

Too much magnesium can mainly be harmful, especially for people with heart, digestive or kidney problems.
Common side effects of magnesium supplements include diarrhea, nausea, abdominal cramps, rinsing, headache and muscle weakness. In people with kidney problems – that cannot be excreted excessly magnesium – the mineral can accumulate to dangerous levels, causing low blood pressure, problems with urination, confusion, breathing difficulties, an irregular or slow heart rate and even cardiac arrest.
With the aim of giving their advertisements more legitimacy, some AI use to make medical experts seem to make positive claims about magnesium supplements, a practice that Chen called ‘dangerous and unethical’.
“Everyone wants simple and fast,” he said. “But when it comes to serious health needs, there is no quick solution for good sleep, exercise and good nutrition.”

The search for a speed of a long service life has led to a tree in the supplement industry. In 2024, the market for nutritional supplements was worth $ 189 billion. According to Market Research Agency Precunest Research, that number is expected to reach $ 402 billion.
To the demand for advertisements with medical testimonies and questionable health claims generated by AI, the social media platforms offered limited statements of their policy. A YouTube spokesperson pointed to his guidelines for the information about wrong information that prohibit content that contradicts the guidelines for the local health authority, such as promoting harmful alternative treatments or discouraging professional medical care.
A representative for Meta also refused to comment directly on the advertisements, but be on their advertising policy, which prohibits misleading or misleading health claims, promotion of unsafe supplements, overly health-related promises and advertisements with negative self-perception tactics or use of medicals.
Tiktok did not respond to Decrypts Request for comments.
“People love these platforms for freedom of expression and the ability to say what they want,” said Chen. “But we need an obligation for the patient’s safety, ensuring that people do not pursue something that is not only not useful, but can also ensure that they ignore a serious condition. If they postpone the search for help, it can be too late.”
According to Dr. Zhaoping Li, director of the UCLA Center for Human Nutrition, while magnesium is now often promoted for sleeping and muscle support, the wider use of it comes from a more specific medical application.
“For women with pre -eclampsia it was used to relax womb muscles during contractions and early birth,” Li said Decrypt. “From there, people started to use it wider for muscle relaxation. Although not extensively studied, small studies suggest benefits for restless Beens syndrome and for relaxing muscles before bedtime.”
Magnesium deficiency is common in people with bad diets or chronic alcohol consumption, Li explained, therefore it is often supplemented in clinical environments.
“If you came to first aid, they would have given you important nutrients – Magnesium is one of them,” she said. “Based on that background, people now recommend it more freely for various health benefits, but the real advantage depends on what you use it for.”
Chen and Li both said that certain forms of magnesium are better absorbed and can be useful in specific cases. But products that are labeled as “complexes” with a dozen types are largely marketing.
“If we really do it for everyone’s advantage, I would recommend taking more natural food with a lot of magnesium,” Li said. “The advantage is much, much more than taking a lot of pills.”
Li argues that the real promise of AI in health is not in marketing supplements, but in understanding food.
“If we really want to use AI, we have to study food completely,” she said. “Do not insulate a mineral and throw it in a capsule.”
Generally intelligent Newsletter
A weekly AI trip told by Gen, a generative AI model.