A newsletter on Responsible AI and Emerging Tech for Humanitarians 

When we talk about AI in humanitarian work, much of the attention rightly goes to crisis prediction, early warning systems, or even direct service delivery to affected communities. But there’s a less visible, yet equally critical, layer of humanitarian response where AI is quietly starting to reshape how we work: the operational support functions that underpin emergency response. 

Operational support functions include the logistical and supply chain functions that keep humanitarian response moving, from procurement and inventory management, to warehousing, transportation, and final-mile distribution. These "back office" systems form the backbone of every humanitarian mission, ensuring that food, shelter, medicine, and other vital supplies reach those in need, often under enormous pressure and in highly unstable environments. 

AI offers new opportunities to strengthen these systems, for example: 

• Predictive analytics can forecast supply chain disruptions, 
• Machine learning models can optimise warehouse inventory and routing, 
• Natural language processing (NLP) can facilitate analysis and knowledge management to supports operations. 

AI can also help monitor evolving risks, such as geopolitical instability, pandemics, natural disasters or even conflict indicators themselves, that threaten global supply chains and humanitarian logistics.

But alongside these opportunities, risks must not be overlooked. Many humanitarian organisations face limited capacity to implement, maintain, and govern complex AI systems. Data used to train models may reflect global inequalities or bias, and automation may create dangerous blind spots if not carefully monitored. In fragile settings, introducing AI too quickly, without strong human oversight, risks undermining the very communities that humanitarian response aims to serve.

As we explore AI's role in humanitarian operational support functions, the question is not only what can be automated, but how we ensure these tools are used responsibly, ethically, and with the flexibility that unpredictable crises demand. 

Case Study: AI powered knowledge retrieval chatbot – Norwegian Refugee Council (NRC)

The Norwegian Refugee Council (NRC) has developed an AI-powered knowledge retrieval chatbot to help staff quickly access internal policies and programme guidance, for example the provision of cash assistance.  

NRC are using a custom-trained,small Large Language Model (LLM) based on Llama, which is hosted securely in the organisation's private cloud. The system allows users to engage with organisational knowledge in natural language. This improves efficiency and significantly reduces time spent manually searching documents.  

Built with strong safeguards to prevent hallucinations and tested in multiple languages, the tool highlights how AI can strengthen humanitarian operations behind the scenes, while careful human oversight remains essential to ensure accuracy and relevance. 

📎 Get in touch: Zineb.bhaby@nrc.no

Read the full case study

Who’s Doing What 

A snapshot of promising AI tools being used across the humanitarian sector. 

1) UNICEF – The world’s largest humanitarian warehouse
UNICEF’s humanitarian warehouse, the largest in the world, has evolved over decades using advanced logistics, data-driven forecasting, and a robot-assisted system.

📎 Contact: supply@unicef.org

2) The World Food Programme (WFP) – Optimus: data driven insights to save lives
Optimus is the WFP’s AI-powered optimisation tool that helps humanitarian operations plan food baskets, develop sourcing strategies, and chose distribution networks.

📎 Contact: global.innovation@wfp.org

3) Macro-Eyes - Developing predictive supply-chains using machine learning for improved immunization coverage
UNICEF grantee Macro-Eyes uses predictive supply chain for vaccines to forecast demand, optimise stock levels, and reduce both shortages and wastage, which is ultimately improving immunisation coverage.

📎 Contact: innovation@unicef.org

Explore the full Humanitarian AI Directory

Tech Spotlight

The AI Lighthouse – The guide to responsible AI for nonprofits 

Navigating AI in the nonprofit world just got easier. NetHope’s AI Lighthouse is a dynamic new resource hub designed specifically for nonprofits looking to implement AI responsibly and purposefully.

Whether you are just starting or scaling your AI journey, the Lighthouse offers curated tools, case studies, and frameworks grounded in ethical best practices. From data governance to lessons learnt from real-world pilots, it's a ‘one-stop shop’ to help humanitarians harness AI's potential without losing sight of their values or the risks.

Explore the AI Lighthouse

Editor’s Choice 

Curated reads and resources our team found especially insightful this month. 

📖 Who Writes the Code?, J. Scurell & M. Mircovic (2025) As AI takes on a growing role in humanitarian operations, from logistics optimisation to service delivery, this piece explores critical questions of who controls AI design, how operational decisions are being automated, and what this means for power, accountability, and equity in humanitarian response. 

📖 Shared Intelligence for Resilient Supply Systems, World Economic Forum (2023) explores how the Global Supply Resilience Initiative (GSRI) leverages real-time data sharing between organisations to make global supply chains stronger and better able to handle crises, using a UNICEF project as an example.

📖 Ai-Driven Logistics for Humanitarian Supply Chains, Prism Sustainability Directory (2025), explores two futures for AI in humanitarian logistics: 1.  Atrophy, where poor oversight worsens inequalities, and 2. Ascend, where responsible AI transforms operations and empowers communities.

📖 How to manage AI risks: a community-based approach to AI assurance, Nesta (2025), explains how involving communities directly in shaping and reviewing AI tools can make humanitarian operations safer, more trusted, and better aligned with local needs.

Skill Up 

Short, practical learning picks for practitioners - no tech background needed. 

Humanitarian Leadership Academy’s AI & Humanitarians: what are the implications for skills and people development in the sector? (free) - Nesta and Nethope discuss AI skills needs across humanitarian organisations, including leadership, technical, and operational roles. 

Community Crisis Intelligence (free) - Now available in Spanish here! -Tools and practices that combine data from affected communities and responders with AI for more effective crisis mitigation, response or recovery – recommended by 93% of participants. This course was created by the Centre for Collective Intelligence at Nesta, in collaboration with Data Friendly Space, Humanitarian Leadership Academy and CDAC. 

Voices from the sector on emerging tech deployment in humanitarian response. 

Humanitarian Frontiers in AI, episode 6 ‘Nowhere To Go but Up: Future Trends of AI Use in the Humanitarian Sector’ is hosted by Chris Hoffman (Tech Journalist) and Nasim Motalebi (World Food Programme) with special guests Nana Gamkrelidze, Planning, Monitoring, Evaluation, and Reporting Delegate at IFRC and Karin Maasel, CEO at Data Friendly Space. 

This episode explores how AI agents can improve internal processes and break down organisational silos, including automation of data collection and analysis which are all closely tied to back-office and operational efficiency. 

🕐 Run time: ~45 min

Stay ahead of funding calls and events. 

UKHIH Humanitarian Responsive Fund - Deadline: 27 June 2025 A grant of up to £60,000 for pilot innovations, research, tools, or events that address emerging humanitarian crises (e.g., Gaza, Sudan, DRC, Haiti, Myanmar, Ukraine).

ARIA Safeguarded AI Research Call – Deadline: 17 August 2025 Funding for research into mathematical techniques that ensure AI systems are safe, reliable, and verifiable - relevant for humanitarian tech where trust and accountability are critical. Open to researchers and consortia (including NGO-academic partnerships). 📎 More info

15th RDR Annual Conference - Where is the ‘Human’ in Humanitarianism? - When: 1 July 2025, London (free, in person). A day of panels and workshops exploring human-centred approaches to disaster risk reduction and humanitarian response amidst digital innovation and geopolitical change. 📎 More info

Unblocking Development: How blockchain can promote inclusive finance – When: 2 July 2025, Seville, Spain. This event explores how blockchain technology can enable faster, cheaper, and more transparent financial transactions  in humanitarian and development contexts. 📎 More info

Unblocking Development: How blockchain can promote inclusive finance – When: 4-5 August 2025, Washington DC Brings together global policymakers, practitioners, and innovators- including digital aid solutions - to explore humanitarian aid, peace and service delivery strategies. 📎 More info

Disclaimer: The views expressed in the articles featured in this newsletter are solely those of the individual authors and do not reflect the official stance of the editorial team, any affiliated organisations or donors.