The personal nutrition revolution: AI, wearables and the future of eating

We assess whether new artificial intelligence models could help or hinder the future of diet tracking. 

We’re used to our phones knowing everything, from our sleep, steps, stress levels, and even our heart rhythms. But the next frontier in tech isn’t about tracking what your body does, it’s about deciding what your body needs before it happens. According to a recent preprint,, scientists have developed a large-language-model system capable of creating personalised meal plans with remarkably low error rates. In other words, AI can now design your diet. It is faster, cheaper, and maybe even smarter than a human nutritionist.

It’s the latest step in what online wellness experts are calling the “personal nutrition revolution,” where algorithms, wearables, and genetic data are merging to take the guesswork out of eating. We’re moving from “What should I eat?” to “What should I eat?”. But is it a shift that promises better health, or does it raise big questions about data, bias, and the meaning of healthy living in an AI-driven world?

For decades, nutrition advice has been contradictory: carbs are bad, fats are good, or maybe it’s the other way around. Even trained experts disagree. But AI thrives on information, so the more data, the better. The new LLM-based system, trained on thousands of datasets, can instantly generate meal plans aligned with medical needs, allergies, budgets, and cultural preferences, while maintaining nutritional balance. Pretty neat, huh?

Unlike standard apps that rely on calorie counting or preset menus, this model can factor in everything from gut microbiome data to glucose-response trends from continuous glucose monitors. Combined with wearable data that is already taking the wellness sphere by storm (heart-rate variability, sleep quality, stress metrics), it can predict how your body might respond to a specific meal before you eat it.

It’s a vision that sounds futuristic, but it’s already here in fragments. Tools like Zoe, Nutrisense, and Lumen are building early versions of this system. The difference now is that LLMs can reason contextually. It will not just crunch numbers but interpret nuance, such as why someone might skip breakfast or crave late-night snacks. It’s a rather humanistic approach for a set of data and numbers.

For people with conditions like diabetes, heart disease, or IBS, this kind of tailoring could be life-changing. Early trials show that personalised AI meal plans improve outcomes, particularly when they integrate behavioural coaching. Some models even account for cultural eating patterns by swapping quinoa for rice or offering halal or vegetarian equivalents, which could finally see a shift in breaking away from the Western-centric bias in traditional nutrition databases – a growing issue when it comes to the wellness industry.

And for Gen Z, the first generation to grow up with both food guilt and fitness trackers fighting against each other, AI may finally provide clarity.

But before we hand over our meal plans to machines, it’s worth asking: what happens when we sell our eating habits to algorithms? AI systems are only as good as the data they’re trained on: and nutrition science is steeped in inconsistencies, socioeconomic gaps, and Western bias. If the datasets used don’t reflect diverse populations, the advice could still marginalise groups with different cultural diets or access to ingredients. Then there’s privacy – an increasing issue when it comes to selling our biological data. Feeding AI systems real-time data from your smartwatch, grocery receipts, and gut microbiome raises major ethical concerns. Who owns that data? 

There’s also the psychological side, where food anxiety consumes our daily lives. If we left our decisions to an app, we would lose the intuition that comes with hunger, hormones, or tiredness. That echoes what’s already happening on TikTok, where nutrition tracking and body data have turned performative. Adding AI into that ecosystem could blur the line between empowerment and obsession.

The most balanced view is to see AI as an assistant, not a director of your diet. Hybrid models are emerging that pair AI analysis with human expertise. Apps like Zoe still use AI to process microbiome data, but human dietitians review the results. This blend of empathy and efficiency could define the future of nutrition care.

We’re standing at the edge of a revolution, not just in how we eat, but how we understand ourselves through food. The same way Spotify knows your musical moods (alarmingly well, may we add) and Netflix your cinematography tastes, your nutrition AI might soon know your cravings before you do.