September 3, 2026·14min·Society·– views

From Being Able to Eat, to Being Able to Choose

Zero Hunger is not the end but the floor. A proposal for the Food Choice Index: measuring not whether people eat, but what they can choose to eat — and using AI to find the bottleneck that limits human food choice.

As of: September 3, 2026

Prologue — Being Able to Eat When Hungry

For some people, food is too ordinary to ever become a question.

When hungry, they eat. When they want something, they pick it. Rice today, noodles tomorrow. Meat some days, vegetables other days. They can eat near home, order delivery, or go to the market. Even if they cannot eat the most expensive food every day, the very range of what they can eat is wide.

For this person, the problem of food is usually not the existence of food. The problem may be price, time, health, or taste. But the basic premise remains.

When hungry, they can eat.

Seen against the whole of human history, this is by no means an obvious state.

Even in 2025, about 645 million people experienced hunger, and about 2.1 billion experienced moderate or severe food insecurity. About 2.69 billion could not afford a healthy diet.

FAO SOFI 2026

But here one more question arises.

Is everyone who does not starve eating enough?

And there is a deeper question.

Can everyone who eats enough choose what they will eat?

This question does not replace the question of hunger. Rather, it pushes that question all the way to its end.

When we think about the food problem, we easily imagine two worlds. On one side the starving; on the other, the affluent.

Reality is not so binary.

A person can experience food insecurity without starving. They can consume enough calories yet lack the means to buy a healthy diet. They can buy food, yet their options can be far too narrow. Food can exist in the market yet be physically hard to reach. They can afford it today, yet be unable to make the same choice next week if prices rise.

The FAO definition of food security explains that all people must at all times have physical and economic access to sufficient, safe, and nutritious food for an active and healthy life, and that the food must meet their dietary needs and food preferences. The core dimensions of food security are availability, access, utilization, and stability, and the recent FAO conceptual framework also emphasizes agency and sustainability.

Sources: - FAO, An Introduction to the Basic Concepts of Food Security: https://www.fao.org/fileadmin/user_upload/eufao-fsi4dm/docs/concepts_guide.pdf FAO AGROVOC, Food Security: https://agrovoc.fao.org/browse/agrovoc/en/page/?clang=en&uri=c_10967

FAO, Food security concepts and measurement: https://www.fao.org/4/y4671e/y4671e06.htm

So the idea of choice is not a new philosophy unrelated to food security. Rather, food preferences are already inside the existing definition.

Then the question is this.

Why do our measurements and policies not measure that power of choice directly enough?

The major indicators of the food problem today each answer an important question.

Hunger indicators tell us how many people fail to obtain, over a long period, the energy they need.

The Food Insecurity Experience Scale shows the degree to which people experience difficulty in securing enough food reliably.

The cost of a healthy diet shows the minimum cost of composing a healthy diet and the population that cannot afford it.

All of these are necessary.

But the actual human experience of food has one more layer.

The space of choosable food.

Suppose two people both eat 2,000 kcal a day.

One in effect eats only a few cheap foods over and over. The other chooses the same calories from dozens of foods and meals.

Both can be classified as “not starving.”

But the food environments they live in are not identical.

For one, food is the minimum survivable set; for the other, food is a wide choosable set.

Dietary diversity is an important principle of a healthy diet. The WHO names diversity as a core principle of a healthy diet and recommends, among other things, at least 400g of fruits and vegetables a day for those aged 10 and over.

Sources: - WHO, Healthy diet: https://www.who.int/en/news-room/fact-sheets/detail/healthy-diet - WHO, Healthy diet --- health topic: https://www.who.int/health-topics/healthy-diet

So the range of food options is not a mere matter of taste.

The breadth of choice is connected to the breadth of nutrition.

Suppose you want steak today but cannot eat it every day because of the price. That is not the same thing as a “failure of food access.”

Conversely, if you are hungry today and skip a meal because you have no money, the story is entirely different.

The first person experiences the price of choice. The second experiences the collapse of choice.

So the food policy of the future needs to see these two separately.

We do not need to provide everyone with every luxury food.

But at minimum we can ask the following.

• Can a person eat when hungry?

• Can they eat healthily?

• Does price narrow their healthy options too far?

• Can they choose food that fits their culture and taste?

• Will that power of choice still hold tomorrow?

Put these questions together and you get what we call Food Choice.

If we describe the world food problem only as “the planet lacks food,” we miss much of the problem.

The World Bank explains that despite record food production, high transport costs in poorly connected and fragile regions can worsen food insecurity, and that in some cases logistics costs can account for half the final food price.

Sources: - World Bank, Increasing Connectivity for Enhanced Food Supply Chain Resilience: https://www.worldbank.org/en/results/2024/04/11/increasing-connectivity-for-enhanced-food-supply-chain-resilience

Production and access are not the same thing.

The fact that food was produced on a farm does not guarantee that it reaches the consumer at a reasonable price.

In between lies a long system: harvest, sorting, storage, refrigeration, processing, transport, wholesale, retail.

And a bottleneck at any one point shrinks the consumer’s choice.

Think of the cold chain as a mere convenience and you underestimate its meaning.

The FAO explains that insufficient refrigeration may cause the loss or waste of about 526 million tonnes of food a year — about 12% of the world’s food.

Sources: - FAO, Cooling the chain, cutting the waste: https://www.fao.org/energy/news-and-events/news/news-details/cooling-the-chain--cutting-the-waste/en

For fresh food especially, temperature and time are critical.

In the FAO’s SDG food-loss data, the global food loss rate for fruits and vegetables in 2023 was about “25.4%,” and the overall global food loss rate — across post-harvest farm, transport, storage, wholesale, and processing stages — was estimated at about “13.3%.”

Sources: - FAO SDG Indicator 12.3.1: https://www.fao.org/sustainable-development-goals-data-portal/data/indicators/1231-global-food-losses/en

So improving cold-chain infrastructure can, in some regions, lead directly to:

more fresh food → less loss → more stable supply → more choice

— that chain.

According to UNEP, about 1.05 billion tonnes of food waste were generated worldwide in 2022. That is nearly one fifth of the food available to consumers, with households accounting for 60%, food service 28%, and retail 12%.

Sources: - UNEP, World squanders over 1 billion meals a day: https://www.unep.org/news-and-stories/press-release/world-squanders-over-1-billion-meals-day-un-report

These numbers do not mean that all discarded food could simply be sent to the hungry. Location before disposal, safety, time, logistics costs, quality, and demand all differ.

But one fact is clear.

The world food system contains enormous inefficiency.

So the food problem is a problem of production volume and, at the same time, a problem of system efficiency.

In 2025 the world average cost of a healthy diet was 4.28 PPP dollars per person per day. It was 3.44 PPP dollars in 2021 and 2.94 PPP dollars in 2017.

In 2025, about 2.69 billion people — “32.7%” of the world’s population — could not afford a healthy diet. In Africa the share was “66.6%.”

Sources: - FAO SOFI 2026: https://www.fao.org/newsroom/detail/un-report--global-hunger-levels-ease-for-third-consecutive-year-as-regional-disparities-persist/en

This creates an important turning point.

We must now ask not only “do people eat” but “what food can they eat.”

And the cost of a healthy diet is not determined by a single food price.

Agricultural productivity, supply chains, market access, logistics, refrigeration, energy, income, and policy all work together.

The food problem can be layered as follows.

Level 1 --- Survival

Can you eat today?

Level 2 --- Sufficiency

Can you continuously obtain the energy and nutrition you need?

Level 3 --- Health

Can you afford a healthy diet?

Level 4 --- Accessibility

Can you obtain the food you want within reasonable time and distance?

Level 5 --- Diversity

Can you choose from a sufficiently diverse range of foods?

Level 6 --- Stability

Does that power of choice hold even under price, climate, conflict, and supply-chain shocks?

This is not a proposal to discard the existing food-security framework.

It is a proposal to measure agency and diversity more precisely, on top of the existing availability, access, utilization, and stability.

“More choice is good” alone lacks policy value.

Why does the power of choice matter?

Food is a good necessary for survival and, at the same time, a cultural, social, and personal choice.

For some people rice matters; for others, bread. In some cultures particular grains and legumes are central; in others, fish and vegetables.

The WHO likewise explains that the composition of a healthy diet can vary with individual characteristics, cultural context, and the foods and eating habits available locally.

Sources: - WHO Europe, Nutrition for a healthy life: https://www.who.int/europe/news-room/fact-sheets/item/nutrition---maintaining-a-healthy-lifestyle

So Food Choice is not “the consumption of luxury food.”

It is:

health + culture + personal preference + economic choice

— the intersection of these.

This report proposes the **Food Choice Index (FCI)** as a new research frame.

It is not an international indicator currently adopted officially by FAO, WHO, the World Bank, or the UN.

The FCI must include at least the following.

Availability

Does the food exist?

Affordability

Can it actually be purchased?

Accessibility

Can it be physically reached?

Diversity

Are there enough kinds of food to choose from?

Healthfulness

Is the share of healthy foods among the options sufficient?

Reliability

Does that power of choice hold over time?

Agency

Does the individual have real freedom to choose according to their culture and preference?

The intuition that “fifty or more food options one wants to eat” is interesting, but the number 50 is not an internationally validated standard at present.

So it must be studied like this.

H10 / H25 / H50 / H100 / H250

For people with each of these choice sets, compare their:

• dietary diversity

• nutritional adequacy

• food satisfaction

• food insecurity

• food expenditure

• food waste

• health outcomes

— compare these.

The question is:

From what level of real options do meaningful changes appear in the human diet and quality of life?

— that is the question.

To properly measure “fifty kinds,” the unit of choice must be defined.

• Ingredient --- rice, potatoes, apples

• Food product --- soy milk, bread, frozen foods

• Dish --- kimchi stew, pasta, curry

• Meal --- an actual meal

• Nutritional option --- nutritionally independent combinations

It depends on the purpose of the research, but the core of the FCI is likely most useful when centered on real choice at the level of meals.

The food system is too complex.

You must not look only at farms.

Not only at prices.

Not only at logistics.

Not only at income.

Not only at nutrition.

Not only at weather.

Everything must be connected.

AI can integrate the following data.

• agricultural production

• inventories

• food prices

• roads

• ports

• refrigeration facilities

• electricity

• income

• population

• food consumption

• nutrition

• weather

• climate risk

• supply-chain shocks

And it can ask the following question.

“What is the biggest bottleneck limiting food choice in this region?”

The traditional approach:

Food is scarce → increase production.

The new approach:

Choice is low → why is it low? → what is the biggest bottleneck? → which intervention

restores the most choice at the lowest cost?

In region A, production may be the problem.

In region B, refrigeration may be the problem.

In region C, roads may be the problem.

In region D, purchasing power may be the problem.

In region E, the price structure of healthy food may be the problem.

Phenomena that look like the same ‘food shortage’ have different causes.

The goal must not be to replicate the current high-income-country consumer’s diet for everyone.

According to the FAO, the world’s agrifood systems emitted about 16.5 billion tonnes of CO2e in 2023, about **32%** of global emissions.

Sources: - FAO, Greenhouse gas emissions from agrifood systems: https://www.fao.org/statistics/highlights-archive/highlights-detail/greenhouse-gas-emissions-from-agrifood-systems.-global--regional-and-country-trends--2001-2023/en

So the goal must be:

more choice

and at the same time

less waste of resources

— both.

Conceptually:

[

\max

{=tex}; Sustainable Human Food Choice ]

subject to:

• nutrition standards

• budget

• land

• energy

• carbon

• supply-chain stability

• cultural fit

— these constraints.

That is:

With limited resources, how do we create the greatest possible human food choice?

AI’s role is fivefold.

1. Measure

Measure current Food Choice.

2. Diagnose

Find the causes where it is low.

3. Predict

Predict how it changes when shocks occur.

4. Simulate

Simulate the results of applying policies.

5. Optimize

Find the most effective combination of interventions within a limited budget.

For each intervention (j):

[ ROI_j = rac{

\Delta

{=tex}FCI_j}{Cost_j} ]

— calculate this.

Including environmental costs:

[ AdjustedROI_j = rac{

\Delta

{=tex}FCI_j} {Cost_j +

EnvironmentalCost_j} ]

This changes the policy question.

Old:

Where do we put more money?

The new question:

Where can each dollar restore the most human food choice?

In the long run, regional Food Choice Digital Twins can be built.

Inputs:

Supply

production, crops, livestock, fisheries, inventories

Logistics

roads, transport, warehouses, refrigeration, ports

Markets

prices, supply, demand, competition

People

population, income, food expenditure, consumption

Nutrition

nutrients, cost of a healthy diet, diversity

Risk

climate, conflict, trade shocks, supply-chain disruption

Outputs:

• current FCI

• bottlenecks

• effects by policy

• risk scenarios

• investment priorities

Existing maps mostly show hunger and food insecurity.

The new map overlays the following.

Layer 1 --- Hunger

Who is starving?

Layer 2 --- Food Security

Who eats reliably?

Layer 3 --- Healthy Diet

Who can eat healthily?

Layer 4 --- Diversity

Who can choose diverse foods?

Layer 5 --- Accessibility

Who can reach food at the time and place they want?

Layer 6 --- Reliability

Whose choice collapses in a crisis?

This lets us see the food problem at the scale of human daily life.

Simulate the following policies in the same region.

Scenario A

Refrigeration infrastructure +20%

Scenario B

Fruit and vegetable prices -10%

Scenario C

Low-income purchasing power +10%

Scenario D

Rural road improvement

Scenario E

A+B+C+D applied together

In each scenario:

[

\Delta

{=tex}FCI ]

— compare this.

Set policy priorities according to the results.

Without claiming these as internationally validated goals, the following are proposed as benchmarks for the research program.

• Establish data standards for computing the FCI in all participating countries

• Build high-resolution Food Choice Maps in at least 100 cities

• Pilot in at least 20 low- and middle-income countries

• Empirically test the H50 hypothesis

• Estimate the FCI effects of refrigeration, logistics, income, and price interventions

The goal is not for everyone to eat the same food.

Food may differ from region to region.

What matters is:

• sufficient nutrition

• health

• diversity

• price accessibility

• cultural fit

• stability

— these.

That is:

Universal Minimum Food Choice

— that is the goal.

AI must not command people what to eat.

The core principle is:

AI optimizes the conditions of choice, not the individual's choice itself.

AI can find ways to lower food prices, raise accessibility, expand healthy options, and stabilize supply chains.

But it must not decide an individual’s taste and choice in their place.

The world is not simply a planet without food.

Even in 2025, 645 million people experienced hunger, 2.1 billion experienced moderate or severe food insecurity, and 2.69 billion could not afford a healthy diet.

At the same time, about 1.05 billion tonnes of food were wasted in 2022; insufficient refrigeration may cause the loss or waste of about 526 million tonnes of food; and the 2023 food loss rate for fruits and vegetables was about 25.4%.

Sources: - FAO SOFI 2026: https://www.fao.org/newsroom/detail/un-report--global-hunger-levels-ease-for-third-consecutive-year-as-regional-disparities-persist/en - UNEP: https://www.unep.org/news-and-stories/press-release/world-squanders-over-1-billion-meals-day-un-report - FAO cold chain: https://www.fao.org/energy/news-and-events/news/news-details/cooling-the-chain--cutting-the-waste/en - FAO food loss: https://www.fao.org/sustainable-development-goals-data-portal/data/indicators/1231-global-food-losses/en

So the next food revolution does not have to be a revolution in production volume.

Innovating the way produced food reaches human beings.

And the stage after that:

Expanding the range of what humans can actually choose among the food that reaches them.

— that is the next stage.

Zero Hunger is not the end but the floor.

After it comes:

Food Security → Healthy Diet → Food Choice → Universal Food Choice

— that sequence.

The final goal is a world in which every human being can reliably choose food that is sufficient, safe, healthy, affordable, culturally appropriate, and sufficiently diverse.

AI’s role is not to decide what that person will eat.

It is to find and remove the bottlenecks in the system that decides what that person can eat.

And that question compresses, in the end, into one.