A food crisis rarely arrives with a bang. It slips in quietly—through smaller bags of rice, thinner meals, and prices that change faster than wages. Food is still there. What disappears first is the ability to afford it.
For lower-income families, food absorbs a large share of everyday spending. Price increases quickly produce smaller meals, poorer diets, and cuts to health or education. This is where food security meets national resilience. A country can protect its borders and digital systems while remaining politically fragile when basic food moves beyond the reach of its poorest citizens.
The danger develops months before its political consequences become visible. Rain arrives late. Heat damages a harvest. Fertilizer shipments are delayed. Conflict closes a shipping route. An exporting country, anxious about domestic inflation, restricts sales abroad.
By the time governments begin emergency procurement, the crisis is already old. The first failure was the inability to connect scattered signals early enough to act.
Stay ahead of the geopolitical week.
MD Briefing delivers expert analysis across five global fronts — the Indo-Pacific, energy, geoeconomics, European security, and the Middle East — every Monday morning. Free.
When the Warning System Disappeared
For four decades, the Famine Early Warning Systems Network, or FEWS NET, helped governments and humanitarian organizations make those connections. It combined satellite imagery, weather data, market prices, crop conditions, conflict analysis, and local observation to anticipate acute food insecurity months ahead.
Then, in late January 2025, its public website went dark.
The interruption followed a stop-work order issued after a political decision in Washington. FEWS NET returned in June 2025, yet the vulnerability was clear: societies facing hunger could lose the system helping them see it approach because of a decision in which they had no meaningful voice.
This should change how the Global South thinks about artificial intelligence.
The Wrong AI Ambition
Much of today’s AI debate is organized around technological prestige. Governments are asked whether they have a national large language model, how many data centers they can attract, or when they will join the frontier race dominated by the United States and China.
Artificial general intelligence has become an imagined finishing line, even for countries still struggling to connect agricultural databases, monitor informal food markets, or anticipate disruptions in basic commodities.
The Global South should remain engaged with frontier AI. Advanced models and computing infrastructure will shape economic power, security, and diplomatic leverage. National AI policy also needs an organizing purpose grounded in the risks already approaching its citizens.
Food security offers the clearest place to begin.
Climate volatility is weakening the reliability of historical agricultural patterns. Geopolitical conflict can disrupt grain, fertilizer, fuel, and shipping routes simultaneously. Domestic politics can magnify those shocks.
In July 2023, India halted exports of non-basmati white rice after retail prices rose and heavy monsoon rains damaged crops. The measure affected roughly half of India’s rice exports. India then supplied more than 40 percent of global rice exports, so a decision aimed at domestic price stability immediately tightened markets for buyers across Asia and Africa. Reuters also noted its political sensitivity ahead of India’s 2024 general election.
The case reveals why food has become a national-security issue. An exporter can absorb a domestic adjustment while import-dependent countries face higher prices, urgent procurement, and fewer alternatives. A distant cabinet decision can reach household kitchens within weeks.
A country may host modern data centers, expand digital payments, and deploy sophisticated AI applications while remaining strategically blind to the shock most capable of destabilizing everyday life.
Proximate-Risk AI
The Global South needs an AI agenda built around proximity to consequence.
Proximate-risk AI refers to systems designed around threats close to a society’s survival, stability, and productive capacity, where earlier intelligence can still alter the outcome.
Proximity is wider than geography. A distant drought becomes proximate when it affects a country’s main grain supplier. A foreign war becomes proximate when it disrupts fertilizer or shipping. A crop disease becomes nationally consequential when fragmented reporting allows it to spread.
For food security, such systems could combine satellite observations, weather forecasts, farm boundaries, prices, logistics, trade restrictions, and local reports. Their purpose is practical: giving governments enough warning to release reserves, adjust imports, or support farmers before their options narrow.
The current AI economy rarely distributes attention according to social urgency. Capital, computation, and research talent remain concentrated in actors whose incentives shape which problems receive the most capable systems. Automated trading, advertising, and corporate productivity attract far more resources than smallholder agriculture or informal food-price monitoring.
This produces agenda-setting asymmetry. The inequality begins before access to a model or participation in its governance. It begins when research agendas, datasets, and computing budgets determine which risks become visible to machines.
This adds an upstream dimension to the double asymmetry I have described elsewhere (see: https://medium.com/@yourusername/your-article-link-here). Developing countries already face unequal access to technological capability and limited influence over its rules. They also possess less power to decide which threats enter the AI agenda. That power shapes datasets, investments, and the time institutions receive to prepare.
Food security also reveals the South’s hidden advantage. Developing countries hold much of the ground truth needed to understand their own food systems: planting calendars, farm conditions, local prices, transport routes, and the interaction between climate, conflict, and livelihoods.
This knowledge remains fragmented across ministries, universities, businesses, and communities. AI becomes useful when those fragments are organized: satellite imagery may detect changing vegetation, while local reporting explains that a damaged bridge closed the route to market.
ASEAN Already Has the Pieces
Southeast Asia offers a practical place to build this model.
ASEAN already has the ASEAN Food Security Information System, or AFSIS. Its database covers five major crops and includes production, yields, prices, trade, stocks, and damaged areas; it also produces commodity outlooks and early-warning reports.
The region also has the ASEAN Plus Three Emergency Rice Reserve, or APTERR, a formal mechanism through which ASEAN members, China, Japan, and South Korea maintain rice stocks for emergencies.
AFSIS helps the region see emerging pressure. APTERR gives it a physical response mechanism. The weak point lies in the movement from warning to trusted collective action.
The ASEAN Digital Economy Framework Agreement, or DEFA, creates an opening. ASEAN concluded negotiations in May 2026 with the stated vision of a digitally integrated, secure, and interoperable regional economy. DEFA’s digital rails could support common food data standards, selected cross-border alerts, paperless emergency procurement, and faster settlement for food assistance.
Interoperability still leaves a question of trust. Governments need to know which agency issued an alert, when it was revised, which model produced it, and whether the record changed after publication.
A permissioned blockchain, or a comparable distributed-verification mechanism, could provide that narrow trust layer. National agencies would retain their raw datasets. Cryptographic records would verify the issuing authority, timestamp, model version, validation status, and revision history. NIST describes blockchain as a shared, tamper-evident, and tamper-resistant ledger, a feature suited to cross-border provenance and audit trails.
The policy architecture is straightforward:
AFSIS as the sensing layer. DEFA as the interoperability layer. Blockchain as the verification layer. APTERR as the response layer.
This connects existing institutions. Data quality would still depend on common definitions, capable agencies, and honest reporting, while the verification layer would make critical forecasts traceable across borders.
It would also give DEFA a public purpose beyond digital commerce. Regional integration should help ASEAN survive physical and geopolitical shocks, especially when food-price pressure reaches low-income households first.
A Model for the Global South
ASEAN’s architecture could offer a practical template for wider Global South cooperation.
A digital non-aligned movement could begin with shared capabilities around concrete risks: common alert standards, jointly evaluated models, federated data arrangements, and scenario exercises for droughts, export restrictions, or logistics disruption.
Strategic agency comes from enough domestic and regional capability to maintain, evaluate, and redirect critical systems when a donor withdraws, a vendor changes course, or political priorities shift elsewhere.
At the national level, food-risk AI should be treated as part of a multiplex digital ecosystem. Agricultural authorities, meteorological agencies, statistical offices, logistics firms, universities, local governments, and farmer organizations each hold part of the picture. The state’s role is to orchestrate those fragments, establish common protocols, and prevent technology for real impact from becoming another single-vendor dependency.
The most valuable low-hanging fruit may be unglamorous: improving local reporting, connecting existing databases, agreeing on definitions, and training officials to challenge model outputs. These steps build institutional muscle that can later support early warning for floods, zoonotic disease, energy disruption, or climate-induced migration.
Proximate-risk AI is an entry point into anticipatory governance.
The Intelligence to See Our Own Future
The Global South must define technological ambition through the risks it is meant to address.
Frontier models will continue to shape power. Computers will remain strategically valuable. Countries also need intelligence systems that protect their room to act before climate, conflict, or political decisions elsewhere become domestic emergencies.
FEWS NET eventually returned. The interruption had already revealed how quickly strategic visibility could disappear.
The defining AI question for the Global South concerns the systemic shocks closest to human survival, and the capacity to model them before outside decisions determine the available response.
In the emerging AI order, sovereignty may be measured less by the size of a country’s model than by something more fundamental: whether it can see its own shocks approaching before the poorest citizens are forced to absorb them.
