Transforming Agriculture: AI for Resilient and Inclusive Food Systems

Contents

Executive Summary

This OECD-hosted panel discussion examines how AI can strengthen global food systems through increased productivity, resilience, and inclusion, while addressing the digital divide that threatens to deepen agricultural inequalities. The session brings together government, industry, and research leaders from the Netherlands, Indonesia, India, and international organizations to explore concrete applications, governance frameworks, and partnerships needed to ensure AI benefits are equitably distributed across smallholder farmers and low-income economies.

Key Takeaways

AI is a powerful tool for agricultural productivity and resilience, but only when embedded in trustworthy, transparent systems with genuine farmer participation and local problem definition. Generic deployment fails; context-driven, farmer-centered design succeeds.
The digital divide in agriculture is widening fast. Without deliberate investment in connectivity, capacity building, and affordable solutions for emerging economies and smallholders, AI-driven productivity gains will deepen existing inequalities rather than democratize opportunity.
Data governance and farmer data ownership must be foundational priorities, not afterthoughts. Fair data infrastructure, transparency in model decisions, and farmer autonomy are prerequisites for adoption and trust—especially in low-tech farming environments.
Effective AI scaling requires problem-specific centers of excellence (e.g., "cold chain resilience," not generic "AI COEs") that align government, industry, and research around measurable outcomes and commercialization pathways.
Anticipatory systems powered by AI can shift food security from reactive crisis management to preemptive response, but require cross-sectoral data sharing, interoperable governance frameworks, and real-time decision-support infrastructure.

Key Topics Covered

AI Applications in Agriculture: Precision farming, crop optimization, pest/disease prediction, weather forecasting, smart irrigation, and logistics optimization
Food System Resilience: Anticipatory action systems, early warning mechanisms, climate adaptation, and shock absorption capacity
Inclusion & Digital Divide: Access gaps between developed and emerging economies; smallholder farmer participation; data governance challenges
AI Governance & Policy: National AI roadmaps, sector-specific regulation, ethical frameworks, and responsible data use
Infrastructure & Capacity Gaps: Digital connectivity, data sharing ecosystems, AI talent distribution, and investment needs
Public-Private Partnerships: Industry alignment, problem-driven collaboration, centers of excellence, and route-to-market strategies
Data Governance & Trust: Data ownership, transparency, farmer autonomy, and ecosystem-wide data infrastructure
Food Security Paradox: Sufficient global food production but persistent hunger, wastage, and distribution problems

Key Points & Insights

Technology Adoption Remains Uneven: Digital tool adoption ranges from 96% of Australian farmers to 12% in Chile, highlighting a critical adoption gap that could worsen existing inequalities if not addressed structurally.
AI Enables Measurable Resource Efficiency: Real-world evidence shows AI-enabled precision spraying reduces pesticide use by 30% without yield loss; computer vision systems cut herbicide use by 50%; smart irrigation saves up to 90% of water; and AI-assisted breeding develops drought-tolerant crops yielding 25% higher yields.
The "Black Box" Problem Undermines Trust: Neural network-based systems lack explainability, which is critical for farmer adoption and regulatory acceptance. Transparent, explainable AI is essential for inclusive agricultural transformation.
Data Scarcity & Ecosystem Fragmentation Limit Impact: Insufficient data sharing agreements, fragmented governance frameworks, and lack of fair data infrastructure prevent models developed at global scale from working effectively at local/smallholder levels—a core barrier to scaling impact.
Problem-Driven Approach is Essential: Generic AI deployment ("throwing AI at every problem") fails. Success requires deep understanding of specific problems (e.g., food wastage in cold chains vs. farmer advisory), local context, and farmer needs before technology application.
Anticipatory Action Framework is Key: AI can shift from reactive crisis management to anticipatory systems that forecast biological and climatic threats, enabling preemptive response before escalation—critical for food security under climate volatility.
Inclusive Ecosystem Approach ("Helix") is Necessary: Indonesia's model demonstrates that successful AI implementation requires collaborative frameworks (government, industry, academia, communities, media) rather than siloed approaches; no single stakeholder can drive sustainable change.
Farmer Engagement Must Drive Data Infrastructure: Rather than treating data as input/output, smallholder farmers must be engaged as active participants in building data infrastructure, owning their data, and understanding how it's used—trust and transparency follow.
Infrastructure Investment is Foundational: Connectivity, digital skills training, affordable solutions, and institutional capacity-building in low-income countries are preconditions for AI adoption—technology alone is insufficient.
Food System Paradox Requires Multidimensional Solutions: Global food production exceeds demand, yet millions face hunger due to geopolitical conflicts, climate shocks, logistics failures, and distribution inefficiencies. AI addresses only specific nodes; systemic barriers require policy, infrastructure, and governance solutions.

Notable Quotes or Statements

"Digitization and AI are powerful tools... They have already proven that they can significantly increase food productivity and reduce food losses." — Ambassador Feder (Netherlands)

"If a farmer or communities are outside of the digital ecosystem, they suddenly are outside of any ecosystem almost. And now with AI it makes it even worse." — Dean Jacob (FAO)

"The biggest problem with AI today is that we throw AI at every problem that exists and we expect that something will happen out of it... If you really want to unlock the technology, you have to know what exactly are you solving for." — Debjani Gosh (NITI Aayog)

"We really need to think how the AI or model which we are developing is applicable to the grassroot level... What we want to solve in the server room or computer doesn't work in the field." — Dr. Aron Pratihast (Wageningen University)

"The key word is anticipation—anticipate the shocks to the agri-food systems that impact food security... Building systems capable of absorbing shocks and implementing anticipatory actions." — Dean Jacob (FAO)

"Inclusive means transparent and explainable... We've had problems with neural network-based systems that are black boxes and cannot be explained in plain [language]." — Professor Arin Sumari (Indonesia)

"While the world is producing enough food to feed 8 billion people, there are still millions who are hungry... There's a paradox; the problem is distribution, access, and the entire ecosystem around it." — Debjani Gosh (NITI Aayog)

Speakers & Organizations Mentioned

Government & Policy Leaders:

Ambassador Feder (Ambassador at Large for AI, Kingdom of the Netherlands)
Professor Arin Sumari (Indonesian Air Force Officer; Professor, State Polytechnic of Malang; Co-chair, Summit Working Group on Economic Growth and Social Good)
Debjani Gosh (Distinguished Fellow & Chief Architect, NITI Aayog Frontier Tech Hub, India; Co-chair, Summit Working Group)

International Organizations:

Dean Jacob (Chief Information Officer & Director, Digital FAO and Agroinformatics Division, FAO/United Nations, Rome)
Sarah Keating (OECD) — Session Moderator
OECD (Organisation for Economic Co-operation and Development)
FAO (Food and Agriculture Organization, UN)
Global Partnership on AI (GPAI)

Research & Academic:

Dr. Aron Pratihast (Senior Researcher, Wageningen University, Environmental Research)
Professor Romesh Chant (NITI Aayog member; mentioned as leading agriculture AI expert)

Companies/Technologies Referenced:

ASML, NXP, Philips (Dutch ICT/tech ecosystem)
European Space Agency (ESA)
NASA Harvest (geospatial data initiative)

Technical Concepts & Resources

AI Applications & Use Cases:

Precision Spraying: 30% reduction in pesticide use without yield loss
Computer Vision for Weed Detection: 50% herbicide reduction
Smart Irrigation: Up to 90% water savings
AI-Assisted Plant Breeding: Drought-tolerant traits identified (sorghum, chickpea); 25% yield increase; shortened breeding cycles
Global AI Hybrid Rice Platform: Predicts optimal parent combinations to enhance resilience
Weather Forecasting & Pest/Disease Prediction: Early warning systems for biological/climatic threats
AI-Enabled Traceability & Logistics: Supply chain transparency, cold chain optimization, transportation route optimization
Crop Mapping: World Cereal Project (ESA partnership) — global cropland mapping with geo-embedding models (NASA Harvest)

Governance & Policy Frameworks:

Indonesia's 7-Pillar AI Roadmap: AI regulation, ethics, investment, data, innovation, talent development, use cases
Quad/Quint/Multi-Helix Model: Collaborative ecosystem framework (government, industry, academia, communities, media)
OECD AI Policy Toolkit: Context-specific guidance for countries; builds on Policy Navigator (oecd.ai), covering 2,000+ policies across 80 jurisdictions
OECD AI Principles: Framework for trustworthy, responsible AI
Global AI Impact Commons: Curated use cases with known impact and scaling potential (summit deliverable)
OECD-GPE Advisory Group on Responsible Agricultural Supply Chains

Data & Infrastructure Concepts:

Data Governance Frameworks: Interoperability standards for cross-border data flows
Farmer Data Ownership: Frameworks ensuring farmer autonomy and data rights
Fair Data Infrastructure: Engagement of smallholders as active data infrastructure participants (not passive input providers)
AI Explainability & Transparency: Requirements for regulatory acceptance and farmer trust

Knowledge & Capacity Building:

Centers of Excellence (CoEs): Problem-specific innovation hubs (e.g., "cold chain resilience," "climate-adaptive crops") vs. generic "AI CoEs"
Farmer Advisory Systems: Phone-based AI services (Government of India tool) enabling multilingual advisory without smartphone requirement
Computer Vision for Disease/Pest Detection: Farmer-centric chatbots with local language understanding

Geographic/Thematic Focus Areas:

Southeast Asia (Indonesia, multi-island challenges)
Central Europe (drought-tolerant crop research)
South Asia (rice cultivation, hybrid rice platforms)
Sub-Saharan Africa & Latin America (smallholder farming systems)
Cocoa agroforestry systems (climate change adaptation)

Document Type: Conference panel discussion transcript
Event: India AI Impact Summit (OECD co-hosted working group session)
Date: Not explicitly stated, but recent (references ChatGPT era, current geopolitical context)
Recording Status: Being recorded; made available on OECD.ai LinkedIn page