How Artificial Intelligence, AI, Can Help Achieve Precision Nutrition

Precision nutrition is about better tailoring diets and dietary recommendations to different people because one size certainly doesn’t fit all, as I’ve written before Forbes. So to determine the best diet for someone you just have to figure out what is going on with genetics, physiology, microbiome, body type, eating behaviors, stress, social influences, food environment, health conditions and all sorts other things. which affects nutrition and health. And you hope how all these things might interact together and change over time. No problem, right?

Not really. Keeping track of these different things happening in different ways at different levels and sorting them out for different people over different times and circumstances can be really complicated. That’s a lot of “different.” These days, however, any time you have something very complex to solve, you have a potential friend in AI – which means artificial intelligence.

One big challenge is that science has not yet figured out how all these different factors might interact to influence how a person’s diet might affect their health. Certainly, studies so far have generated insights into how each of these factors can act separately and for certain types of people. But combining these insights is a different matter and many gaps remain.

That is because a single traditional real-world laboratory, clinical or epidemiological study cannot account for, measure and track the various factors and outcomes for all types of people. No matter how hard you try to design the “perfect” study, you will inevitably fail to include all types of people and measure all relevant factors and outcomes.

Furthermore, even if you were to design the “perfect” study, you would have to wait a long time to get all the results you need. It can take years, even decades for the effects of nutrition to manifest as various health conditions. Anyone who ate like garbage and included ketchup as a vegetable during their 20s will tell you that.

So if you want to figure out how to make a precise diet, you have to combine data from many different studies and fill in the gaps. You also want to find ways to extend the results of a particular study to people who did not participate in that study and circumstances that were not covered. All of this can be too complicated for any one person or even a team of people to do without help.

Enter AI and mute the Randy Newman song, “There’s a Friend in Me.” Such computer-aided techniques can keep track of many different things, combine different data sets in different ways and figure out how they fit together. These techniques can also determine how the results of an individual nutrition study may apply to different conditions and situations, increasing the usefulness and value of that study. And various AI techniques can do it quickly, much faster than humans. These are just some of the ways AI can help achieve precision nutrition.

To understand how AI can do these things, you first need to know what AI is. These days AI has become such a sexy term that people can use it without even really knowing what the term means, like saying things like, “Hey, can you AI this?” AI is an umbrella term that basically encompasses any computer-aided technique that can replicate what the human brain would normally do beyond following step-by-step instructions. So an AI approach could assess situations or make decisions on its own. Many AI approaches, methods and tools already exist and the list continues to grow every year.

One way to classify AI techniques is along a continuum of how these techniques are designed and function. At one extreme is a purely data-driven AI approach. These are “top-down” techniques that start with a body of data and try to figure out patterns, trends and associations from this data. It’s a bit like how a statistician might analyze a data set. But the AI ​​algorithm can do it much faster and perform many different analyzes across multiple data sets at the same time.

Let’s look at a theoretical example. A data-driven AI approach can analyze different data sets, slice the data in different ways and find that people who eat a certain food item tend to live longer. Let’s call this food item “The Best Food Ever”, a completely fictional term named after nothing in particular. The AI ​​algorithm may associate The Best Food Ever with greater longevity but does not explain why this association actually exists. It is not really possible to distinguish whether consumption of the Best Food Ever has some beneficial nutritional effect versus some kind of coincidence. It may be that those who tend to eat the Best Food Ever at the same time as another food item not captured in the data set are actually doing the trick. Or maybe people with less stress are more likely to have the time and money to eat The best food ever. The Best Food Ever could be a red herring, which could mean something misleading or evasive rather than something made from fish.

At the other end of the spectrum are mechanical or interpretable AI approaches. These AI methods try to recreate what is actually happening from the bottom up by recreating the actual mechanisms behind a process or decision. They are considered explainable because you know the specific reasons why a result was generated.

This is similar to what scientists do when they design experiments in a laboratory to test what will happen. The difference is that the AI ​​algorithm or model is not limited to a physical lab and can act as a “virtual lab” representing an entire person, team, population or geographic area. The model can then run experiments in computer “safety” in ways that would be too complicated, too expensive, too time-consuming, too impractical or even too dangerous to do in the real world. The mechanical AI tool could use the results of those experiments to determine recommendations, just as a person runs thought experiments in his or her head before taking action.

So, for example, a mechanistic AI approach could represent the various reasons why a person chooses to eat The Best Food Ever. It could also show the different nutrients in the Best Food Ever, how they are broken down in the body, how these nutrients affect different organs and then how this affects longevity. Then this AI model could look at what would happen over time if different people were to eat The Best Food Ever and decide who would benefit from eating The Best Food Ever and how.

These different AI techniques along the spectrum can also work together and integrate. A purely data-driven approach can suggest associations (eg, take a closer look at The Best Food Ever) which can guide building a more mechanistic AI approach (eg, let’s figure out what which is being done by the best food ever for the body) . Similarly, a mechanistic approach to AI can help define where data-driven approaches are needed. Say when trying to demonstrate the mechanisms by which The Best Food acts on the microbiome but these are impossible to figure out because there are no traditional studies that have clearly shown associations, patterns and trends. Therefore, a data-driven AI approach could be useful for filtering this microbiome data.

Of course, you shouldn’t trust anything AI tells you. A poorly designed clinical trial or observational study can lead to misleading results, and a poorly designed AI approach can lead to misleading results. That’s why you’ve got to know what’s under the hood of AI approaches and understand their relative strengths and weaknesses. At the same time, no AI approach – just like any real-world study – will be perfect. Don’t let the perfect be the enemy of the good and let the imperfections of the AI ​​approach prevent you from using it for risk aversion.

Integrating more AI and other computer-aided approaches to make more accurate recommendations is not entirely new and has been done in other fields. Fields such as meteorology, finance and aerospace engineering have long used computer-aided techniques to bring together and analyze complex data from various sources and generate more accurate insights and predictions.

So while AI is unlikely to challenge some of the already established nutritional insights such as the value of eating fruit and vegetables, the field of nutrition is ripe for change. There are too many people out there claiming that this diet works for everyone. But not everyone is the same and they have the same circumstances, and that is exactly the problem. Achieving more balanced nutrition is not easy. but you have a potential friend in AI. But like any potential mate, you have the right treatment and know what he can and cannot do.