Personalized nutrition with AI: this is what it will look like

The field of nutrition which is perhaps the most timid but is certainly the most important because it concerns the energy we take to live it does not yet seem to have benefited from the digital transformation and in particular from the exponential evolution of AI and connected objects (IoT).

“The digital device you wear on your wrist or on your t-shirt vibrates and lights up, capturing your attention, it informs us that at our doorstep they are about to deliver a protein bar with a chocolate and strawberry flavor accompanied by a fruit and vegetable smoothie, more precisely carrots, strawberries, red apple and ginger with an added dose of vitamin C and E, just how we like it. We didn’t order anything. Or rather, we did not purchase anything directly and consciously, but we gave (ex ante) our consent for everything to happen via an intelligent app. Our order started independently from our device and the AI ​​software with which we monitor and manage our psychophysical well-being. The software chose for us what, how and when; although it would be more appropriate to say that she assisted us in the purchase, and of course she would also be ready to provide all the necessary explanations, if we wanted to know the reasons for this purchase. Ring the bell.” Taken from iFood: how to escape food ideology? (Guerini and Associates).

Investments in agrifood technologies

According to research by AgFunder, «2022 AgFunder AgriFoodTech Investment Report», in 2021 venture capital invested 51.7 billion dollars in agrifood technologies, recording an increase of 85% compared to 2020. The top three by investment are e-grocery (online shops), 35.8% of investments, innovative foods And network and cloud infrastructure (on which service and sales activities are based, such as e-grocery or delivery) with 9.3% of investments.

The e-grocery saw 188% year-over-year growth, but that was thanks to four nine-figure rounds of funding. The cloud-based retail infrastructure saw its funding grow 97.5% to $4.8 billion. Innovative foods received 103% more investment with over 430 companies raising funds. The report makes no reference to investments in app for diet or better nutritiondemonstrating that the market is still scarce or of little interest.

This slowness is not due to the traditional reputation of the food sector and in particular of agrifood which is mistakenly dissociated from technology. The food production sector like the agricultural sector is not as behind the Tech industry as one might think. They are simply more sober and subtle in communicating innovation because the consumer is cautious with everything related to your health.

AI, because in the food sector it is still in the testing phase

In the food equation innovation is often in conflict with tradition which has guaranteed, up to now, health. Therefore, the apparent timidity of food with AI must be interpreted as a running-in phase of a process of radical transformation that will change humanity forever. It should not be forgotten that this is not about a car or other objects and our relationship with them. This is about what we eat to give us energy, joy, well-being, and how we burn calories. It’s about us, about our most intimate psycho-physical state. So while the consumer searches for self-driving cars and any gadget that can make his life easier, he shuns, at least for now, the technology that can tell him what and when to eat.

It should then be highlighted that there are very few food companies with significant organizational and financial capacity sufficient to introduce significant innovations. The sector, especially in Italy, but also in the rest of the planet, is characterized by many medium and small businesses with poor investment capacity and implementation of the most advanced technologies.

The potential of AI in the food industry

In addition to the agricultural sector with smart or precision agriculture, already well argued here on Agendadigitale, the food industry is also about to be overwhelmed by AI. However, it is still difficult to know at the moment the transformations and innovations of the organizational and production processes, products and services to which AI and IoT are contributing.

In the newly published book, iFood, I dedicated myself (chapter 5) to the part that concerns the product they instruments (yes, these are new) to support the consumer in planning their diet, personalizing it. These two factors are functional to overcoming the current – unsuccessful – policies for public health, against obesity. Above all, they are an essential tool to help consumers learn and choose freely, i.e. consciously. Mine is in fact a text that investigates and promotes freedom of choice.

As for the products, there is still little. Worthy of note, more in terms of marketing than innovation strictly speaking, is the new Coca Cola Y3000 whose formula was generated by an AI platform.

On the other hand, presenting billions of images to a machine is complex, but nothing compared to taste, flavor, smells, and colors. The direction taken is undoubtedly this.

Tailor-made food: genetics and nutrigenomics

In the book I call it tailor-made food. Genetics, nutrigenetics and nutrigenomics, similar but distinct concepts, which contribute to producing foods that for example, (i) are more sustainable and resilient, (ii) better resist sudden and violent climate changes, (iii) last longer on the shelf by reducing waste, (iv) they contribute to strengthening food safety, (v) they are more nutritious, tasty, etc., (vi) they respond to individual needs (tailor-made). If there will also be a further (vii) i.e. real time, we will discover it along the way. If we make our own shoes at home with a 3D printer, we will also be able to make customized food. Either the tool will help us choose from the shelf, for example, the vegetables best suited to our needs, or we will be able to produce them at home, to measure, or at least order them exactly as we need them.

Diet apps: between algorithms and health

A series of rudimentary applications already exist which, through the use of simple algorithms, help consumers to choose. They are still rudimentary compared to technological potential. Some limit themselves crudely to presenting the caloric and nutrient content of foods, telling us how healthy they are based on an algorithm that does not include any personal parameters. Others calculate the energy requirements using fixed parameters, such as weight and height, and our food preferences, and variables such as for example the quantity of nutrients we take in at each meal, calorie consumption obtained through heartbeat, movement, etc…

Our body is complex as is the relationship that is established with the energy we take in and the context with which we relate. It is already very complicated to train the software of a car that is affected by thousands of events. Imagine how difficult it can be to train an algorithm that should tell us what to eat to keep us healthy. For these apps that concern health and well-being we would have very high expectations that are difficult to satisfy. The variables that affect our functioning are difficult to categorize and predict. The effect of a teaspoon of sugar taken at this precise moment will have very different consequences on each of us. Our reaction will change if we were to consume those sugars at different times of the day. Furthermore, i factors that influence calorie accumulation there are so many and largely still unexplored. It’s not just a matter of considering the movement we do or our nutrigenomics, but a series of variables that we struggle to imagine or around which the scientific community is divided (see obesogens, for example).

Towards a personalized diet: the importance of data

It’s a matter of time, though. We must be optimistic. The increasingly massive invasion of connected devices (IoT) will allow us to consider and monitor a greater number of psychophysical and external parameters (such as the air we breathe, for example, in the case of obesogens), progressively reducing the number of unknown variables.

The more information we have about our psychophysical state and our relationship with the external environment, the more AI applications will be able to anticipate our needs and develop the most suitable diet that best responds to our most intimate needs. This is a personalized diet. It is not conditioned by external judgments of those who want to establish whether a nutrient is healthy or harmful, but evolves through experience and knowledge of our psychophysical state which evolves over time.

A profound transformation awaits the food sector. It will no longer be about producing food en masse but personalized foods to meet the precise needs of each individual. Industrial and business models will have to adapt. The value will no longer lie in the food itself, but in the process of personalizing food.

A personalized diet means adapting nutrition to the specific needs of each individual, rather than adhering to a universal concept of ‘health’ or ‘healthy eating’. Despite this, uncertainties and concerns remain. In addition to the conservative resistance of those who fear losing traditional culinary pleasures such as mozzarella or local strawberries, food personalization offers the opportunity to go beyond food ideologies and the impositions of one food over another through labels, classification systems such as the Nutri-Score and specific taxes. These methods have shown limits in the fight against obesity, a growing problem as highlighted by recent data.

However, tailoring your diet requires individual awareness and a certain degree of check. It is essential to remain aware that the devices we use are making choices based on extremely accurate, but not infallible, parameters and predictions. Human nature is unpredictable and infinitely complex. Sudden desires, such as those for the chocolate or a steak, can escape the predictions of AI and IoT. These technologies can only mitigate the consequences, suggesting a lower consumption of fats or sugars in case of excesses or encouraging physical activity.

Privacy and security: the challenges of nutrition 4.0

Furthermore, the algorithm behind these systems is always developed by private companies and engineers, based on principles that may be subject to debate and revision. Finally, there is the question of data. In the context of personalized nutrition, we are dealing with extremely personal information, which goes beyond simple vehicle data. We talk about our DNA, our lifestyle habits, in essence, ourselves, including aspects of our consciousness. This raises important questions about privacy and the responsible use of information.