AI for Food Security and Nutrition Innovation Challenge

About the Challenge

The AI4Food Security and Nutrition Innovation Challenge (AI4FSN Challenge) is a nationwide innovation program that mobilizes students, educators, government practitioners, civil society, and the private sector to co-create AI-powered, no-code solutions addressing food insecurity and malnutrition in the Philippines.

The Challenge is designed to break silos—connecting universities, National Government Agencies (NGAs), Local Government Units (LGUs), civil society organizations (CSOs), private sector actors, and startup incubators—to ensure that promising ideas move beyond prototypes toward real-world adoption, incubation, and scale.

Why This Matters

Despite existing food and nutrition programs, gaps remain in:

• Targeting and inclusion of vulnerable households

• Nutrition quality and feeding program delivery

• Coordination across NGAs, LGUs, CSOs, and private partners

• Monitoring, feedback, and last-mile implementation

At the same time, many public servants, educators, and practitioners understand these challenges deeply but lack accessible tools to design AI-enabled solutions.

The AI4FSN Challenge addresses this gap by:

• Enabling cross-sector collaboration, not just student projects

• Making no-code AI tools accessible to non-technical users

• Embedding ethical AI and public-sector safeguards from the start

• Creating a clear pathway to incubation and deployment

Child Nutrition Challenge in the Philippines

Child malnutrition remains a critical and persistent issue in the Philippines. A significant proportion of children under five experience stunting, reflecting chronic undernutrition that affects physical growth, cognitive development, and future productivity. Many children also suffer from micronutrient deficiencies, such as anemia, even in food-producing areas. These challenges are driven not only by poverty, but also by poor diet diversity, limited caregiver nutrition knowledge, gaps in feeding program implementation, and uneven access to health and nutrition services, especially in geographically isolated and disaster-prone communities. Addressing child malnutrition requires better-coordinated, data-informed, and community-responsive solutions—areas where ethical, well-designed AI tools can provide meaningful support.

Who Can Participate

The Challenge welcomes multidisciplinary, cross-sector teams of 3–5 members, including:

Eligible Participants

• Students (undergraduate or graduate)

• Faculty members and researchers

• National Government Agency (NGA) employees

• Local Government Unit (LGU) officials and staff

• Civil Society Organization (CSO) practitioners

• Private sector professionals and social entrepreneurs

Teams are encouraged to mix participants across sectors (e.g., a student, an LGU staff member, and a CSO practitioner).

Note: Government participants join as innovators and domain experts; all solutions remain advisory and augmentative, not automated decision systems.

What Participants Will Build

Teams will design LLM-augmented, no-code AI solutions aligned with one or more focus areas:

• Food Access & Affordability

• Child Nutrition & Feeding Programs

• Social Protection & Food Assistance

• Local Food Systems & Partnerships

• Monitoring, Evaluation & Community Feedback

• Disaster, Climate & Nutrition Resilience

All solutions must:

• Use AI as an augmentation and decision-support tool

• Be feasible using no-code or low-code platforms

• Comply with data privacy, ethical AI, and government safeguards

Challenge Journey

1. National Launch & Onboarding

• National kickoff with AI4Gov Lab and Zero Hunger Lab

• Orientation for universities, NGAs, LGUs, CSOs, and private partners

• Release of problem statements and challenge guidelines

2. Capacity Building

Participants receive training on:

• AI for food security and nutrition

• No-code AI development tools

• Responsible AI, bias, and public-sector safeguards

3. Solution Development

• 8–10 weeks of guided development

• Mentorship from government, academic, civil society, and startup experts

• Focus on usability, feasibility, and real-world relevance

4. Provincial Showcases

• Demo days hosted by partner universities

• Selection of top solutions per province

5. National Demo Day & Incubation Referral

• National showcase and recognition

• Referral of promising solutions to a regional network of business incubators, including:

  • University-based Technology Business Incubators (TBIs)

  • Government-supported innovation hubs

  • Private and social enterprise incubators

From Challenge to Startup or Public Deployment

The AI4FSN Challenge functions as a pre-incubation and deployment pipeline.

High-potential teams may receive:

• Startup or social enterprise incubation

• Support for pilots with NGAs, LGUs, schools, or CSOs

• Guidance on public-sector adoption pathways

• Legal, procurement, and AI governance support

• Access to seed funding, grants, or innovation financing

Solutions may evolve into:

• Startups or social enterprises

• GovTech or CSO-deployed tools

• Open-source or public digital goods

What Participants Gain

For Students

• Real-world AI-for-good experience

• Mentorship and career exposure

• Pathways to incubation and funding

For Government Employees

• Hands-on innovation experience

• Tools to improve program delivery

• Cross-sector collaboration and learning

For Faculty and Researchers

• Applied research and extension opportunities

• Student mentorship and impact pathways

For CSOs and Private Sector

• Co-creation of deployable solutions

• Access to innovation talent and pilots

Our Partners

AI4Gov Lab

Leads governance, ethics, and public-sector alignment to ensure AI solutions are responsible and policy-relevant.

Zero Hunger Lab

Provides technical leadership on food security and nutrition and ensures evidence-based impact.

University & Incubator Network

Supports regional innovation, incubation, and scale-up of high-potential solutions.

Join the Innovation Ecosystem

The AI4Food Security and Nutrition Innovation Challenge brings together government, academia, civil society, and the private sector to co-create ethical AI solutions for hunger and malnutrition.

📌 Universities & Agencies: Nominate teams
📌 Students, Faculty, Practitioners: Apply to participate
📌 CSOs & Private Sector: Co-create and pilot solutions
📌 Incubators: Join the scale-up network

Together, we turn AI ideas into food-secure futures.

Possible Projects

🧭 Beneficiary Targeting & Enrollment

1. EPAHP Eligibility Explainer Chatbot

This project uses an LLM-powered chatbot to translate EPAHP eligibility rules into simple, conversational language. The LLM augments access by explaining criteria, documents, and processes in Filipino or local languages, adapting responses based on user literacy and questions. Built on a no-code chatbot platform, it reduces confusion, misinformation, and unnecessary LGU visits. By clarifying eligibility upfront, the system improves appropriate self-selection, reduces administrative burden on DSWD staff, and ensures that vulnerable households better understand their rights and pathways into EPAHP support.

2. Pre-Screening Intake Assistant

The LLM augments EPAHP intake by guiding households through a conversational pre-screening process before formal validation. Instead of rigid forms, users describe their situation in natural language, which the LLM structures into standardized intake fields. Using no-code tools, this assistant flags likely eligibility, missing information, and next steps without making final determinations. This improves intake efficiency, reduces staff workload, and helps households prepare complete applications, accelerating inclusion of food-insecure families into EPAHP.

3. Household Vulnerability Profiler

This project uses an LLM to synthesize household intake narratives into concise vulnerability profiles for caseworkers. The LLM augments decision-making by summarizing risks related to food insecurity, income instability, health, and household composition in plain language. Using no-code dashboards, staff receive readable summaries rather than raw text. This supports more consistent prioritization, reduces cognitive overload, and helps ensure assistance decisions are informed by the full context of each household’s lived experience.

4. Sectoral Group Mapper

The LLM augments targeting by identifying whether households belong to EPAHP priority sectors (e.g., fisherfolk, IPs, urban poor) based on descriptive inputs. Instead of relying solely on checkboxes, the model interprets narratives about livelihood, location, and cultural context. Built with no-code tools, this helps LGUs more accurately align households with appropriate interventions. It improves inclusivity for groups that may struggle with formal classification while preserving human validation and oversight.

5. Duplicate Assistance Detector (Text-Based)

This project uses an LLM to flag potential overlaps between EPAHP and other assistance programs by analyzing household narratives. The LLM augments coordination by detecting similarities in described benefits, timelines, or providers without requiring full database integration. Using no-code workflows, it highlights possible duplication risks for staff review. This supports more efficient resource allocation, reduces redundancy, and strengthens accountability while keeping final decisions with DSWD personnel.

🍚 Food & Nutrition Program Augmentation

6. EPAHP Feeding Menu Generator

The LLM augments EPAHP feeding programs by generating cost-constrained, nutritionally aligned menus using local food inputs. Through no-code interfaces, implementers specify budget, target population, and available ingredients. The model reasons across nutrition guidelines and constraints to suggest feasible menus, explaining trade-offs in plain language. This improves nutrition quality and consistency across sites without requiring dietitian support for every locality.

7. Nutrition Adequacy Explainer

This project uses an LLM to explain the nutritional rationale behind EPAHP food packages. The LLM augments transparency by translating technical nutrition concepts into caregiver-friendly language. Using no-code chat tools, families learn how food assistance supports child growth and household health. This improves trust, reduces misconceptions, and encourages proper food utilization, strengthening the real-world impact of EPAHP nutrition support.

8. Local Food Substitution Advisor

The LLM augments program resilience by suggesting nutritionally equivalent local food substitutes when standard EPAHP items are unavailable. Through no-code chat, implementers receive alternatives adapted to market conditions and cultural preferences. The model explains why substitutions are acceptable, preserving nutrition quality while maintaining flexibility. This supports continuity of feeding programs amid supply disruptions.

9. Child Nutrition Risk Companion

This assistant uses an LLM to help staff explain child nutrition risks identified during EPAHP engagement. The LLM augments communication by framing risks clearly, empathetically, and without medical jargon. Built on no-code chat tools, it supports consistent messaging, encourages caregiver understanding, and promotes appropriate referrals. This improves early action on malnutrition risks without replacing professional judgment.

10. Feeding Compliance Guide for Partners

The LLM augments partner compliance by translating EPAHP feeding standards into step-by-step operational guidance. Using no-code platforms, NGOs and LGUs receive clear instructions tailored to their role and capacity. The model answers clarification questions in real time, reducing errors and variability in implementation. This strengthens program fidelity across diverse partners.

🏪 Partnerships with CSOs, NGOs & Private Sector

11. Partner Matching Assistant

This project uses an LLM to match LGUs with potential EPAHP partners based on needs and partner profiles. The LLM augments coordination by summarizing compatibility and explaining partnership rationale in plain language. Built with no-code tools, it reduces search and negotiation time, supporting faster, more strategic collaborations.

12. MOA Drafting Assistant

The LLM augments administrative efficiency by generating first-draft MOAs aligned with EPAHP standards. Using no-code document tools, LGUs input partnership details, and the model produces structured drafts for legal review. This reduces drafting time while maintaining compliance and human oversight.

13. Partner Capacity Assessment Bot

This bot uses an LLM to analyze narrative self-assessments from partners. The LLM augments evaluation by summarizing strengths, gaps, and readiness in clear language. No-code deployment enables consistent assessments without complex scoring systems, supporting better partner selection and capacity-building.

14. Private Sector Engagement Pitch Generator

The LLM augments outreach by generating tailored partnership pitches for private sector actors. Using no-code inputs, LGUs receive concise, context-specific messages that align EPAHP goals with corporate interests. This improves engagement quality and reduces staff workload.

15. Partner Feedback Synthesizer

This project uses an LLM to summarize qualitative partner feedback into actionable insights. The LLM augments learning by identifying patterns and recurring issues across narratives. No-code dashboards help DSWD improve partnership management systematically.

🏘️ LGU & Barangay Implementation Support

16. Barangay EPAHP Operations Guide Bot

The LLM augments frontline implementation by acting as a conversational, always-updated EPAHP manual. Using no-code chat, barangay staff can ask procedural questions and receive clear guidance. This reduces dependency on training refreshers and improves consistency.

17. Caseworker Decision Support Assistant

This assistant uses an LLM to summarize household histories and prior interventions. The LLM augments caseworker judgment by presenting key context succinctly, reducing cognitive load and improving continuity of care.

18. Implementation Checklist Generator

The LLM augments operational planning by generating context-specific task checklists. Using no-code tools, implementers receive tailored guidance based on barangay conditions and timelines, reducing missed steps.

19. Inter-Agency Coordination Chatbot

This chatbot uses an LLM to clarify agency roles within EPAHP. The LLM augments coordination by answering questions about mandates and workflows, reducing confusion and delays across agencies.

20. Staff Training Micro-Lesson Generator

The LLM augments capacity building by generating short, role-specific training content. No-code deployment allows rapid updating and scalable staff learning.

📊 Monitoring, Evaluation & Resilience

21. Narrative Progress Report Generator

The LLM augments reporting by converting field notes into structured progress narratives. Using no-code tools, staff save time while improving report quality and consistency.

22. Outcome Story Analyzer

This project uses an LLM to synthesize beneficiary stories into themes. The LLM augments evaluation by capturing qualitative impact often missed by indicators alone.

23. Red Flag Detection Assistant

The LLM augments early warning by identifying risk signals in qualitative updates. No-code deployment supports timely intervention while keeping decisions human-led.

24. EPAHP Indicator Explainer

The LLM augments data quality by explaining indicators and reporting requirements in plain language. This reduces reporting errors and improves compliance.

25. Community Feedback Summarizer

This assistant uses an LLM to analyze open-ended community feedback. The LLM augments responsiveness by identifying trends and concerns quickly.

26. Disaster-Responsive EPAHP Advisor

The LLM augments adaptability by guiding EPAHP adjustments during disasters. Using no-code chat, implementers receive context-aware advice that preserves nutrition outcomes during shocks.

27. Seasonal Hunger Risk Forecaster (Qualitative)

This project uses an LLM to identify emerging hunger risks from narratives. The LLM augments anticipatory action without requiring predictive models or sensors.

28. Urban vs Rural EPAHP Customizer

The LLM augments contextualization by adapting guidance to urban or rural realities. No-code tools ensure implementers receive relevant instructions.

29. Indigenous Peoples Contextualizer

The LLM augments cultural sensitivity by helping tailor EPAHP delivery to IP food practices and norms. This supports inclusive, respectful implementation.

30. Exit & Graduation Readiness Assistant

This assistant uses an LLM to synthesize indicators and narratives to assess readiness for exit from EPAHP. The LLM augments transition planning while keeping decisions with social workers.