A newsletter on Responsible AI and Emerging Tech for Humanitarians 

AI is gaining traction across humanitarian operations, from chatbots that support displaced communities to data that predict food insecurity or disease outbreaks. But as these tools emerge so does a critical issue: not everyone has equal access to them. The rise of AI risks is mirroring, and even amplifying, the same digital divides that have long shaped humanitarian response. 

This month’s edition of Humanitarian AI Unpacked explores the AI divide: between those who have the opportunity to design, deploy, and adapt AI tools and those who don’t. We examine three dimensions of this gap: 

1. Sectoral Divide: Compared to sectors like health or finance, humanitarian organisations lag in adopting AI [1]. Despite digital literacy among humanitarians, only 3.5% report expert AI knowledge and just 1 in 5 organisations has an AI policy in place [2]. Budget constraints, weak data infrastructure, and limited technical capacity are real barriers. While important, partnerships with the private sector risk deepening dependency and raising ethical concerns about data ownership and long-term sustainability [3]. 
2. Global North-South Divide: Most AI tools, policies, and governance frameworks originate in the “Global North”, yet the humanitarian needs they aim to address are overwhelmingly concentrated in the “Global South” [4]. Local NGOs may face steep barriers to access, from limited funding to limited bandwidth and hardware [5]. Southern researchers and practitioners often remain underrepresented in AI model design and high-level policy discussions which risks tools that are misaligned with local realities. 
3. Agency-Community Divide:  Affected communities, particularly displaced people, women, and marginalised groups, are not always involved in the design or oversight of AI systems, even as these tools increasingly influence the services they receive. Too often, “innovation” is done to communities, not with them, reinforcing the very power imbalances humanitarian aid seeks to challenge [6]. 

This edition asks: what will it take to close the AI divide and build a more inclusive and accountable future for humanitarian technology? 

Let’s unpack where we are - and how we move forward. 

Case Study: Closing the forecasting gap with AI

Catholic Relief Services (CRS)

Faced with volatile food prices, Catholic Relief Services (CRS) set out to modernise how its field programmes procure food at best possible prices. The decision on where and when to make food purchases and how much cash is required to cover identified need began as a manual, spreadsheet-heavy task. Food prices however can be impacted by many factors on local, regional and global levels. CRS developed a machine learning-powered dashboard that calculates 6-to-12-month price forecasts which help teams plan cash and voucher assistance, optimise local purchasing, and save costs. 

But while the tech is powerful, the bigger story is about access and usability: 

• Co-designed with the programme teams, the dashboards provide insights for building budgets for food security grants. 
• CRS focused on building tools that non-specialists could actually use - combining complex analytics with local data and practical formats. 
• To address the capacity gap, dashboards are rolled out alongside training, technical support, and clear ownership structures so country teams can adapt and sustain them. 

Key features include: 
✅ Forecasts based on 15+ years of market data 
✅ Tools to flag high-volatility commodities 
✅ Adaptable templates for different modalities (cash, local procurement, etc.) 

Feedback from users is positive: the dashboards save time, increase credibility in budget planning, and enable faster business development. However, measuring procurement savings remains difficult, as tracking this would require time from already overstretched country teams. The team is also addressing infrastructure limitations to ensure the dashboards can scale more widely without compromising performance. 

CRS now aims to scale the tool further but stresses that sustainability depends not just on tech performance, but on ongoing investment in training, ownership, and infrastructure. 

📎 Get in touch: Nora Lindstrom nora.lindstrom@crs.org

Read the full case study

Who’s Doing What 

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

1) CLEAR Global – Multilingual AI for low-connectivity settings
Develops AI-powered language tools (including voice and chatbot interfaces) in underserved languages improving information access for crisis-affected communities with low literacy and limited connectivity. 
📎 Contact: info@clearglobal.org

2) Data Friendly Space – Human Intelligence power by AI
Data Friendly Space’s GANNET is a secure, AI-platform that gives humanitarian workers, especially those in local or capacity-limited teams, real-time access to trusted, multilingual crisis data and analysis to bridge the AI access gap.
📎 Contact: info@gannet.ai

3) American Red Cross – Clara, the Disaster Response Chatbot
“Clara” chatbot uses AI-driven natural language processing to safely guide disaster survivors to shelters, financial aid, blood donation services, and more - extending digital access and trusted support in crisis moments.
📎 Contact: customercare@redcross.org

Explore the full Humanitarian AI Directory

Editor’s Choice 

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

📖 How are humanitarians using artificial intelligence in 2025?, Data Friends Space and Humanitarian Leadership Academy (2025) Based on a global survey finds that while 70% of humanitarians regularly use AI tools, fewer than 22% of organisations have formal AI policies or training - a stark “humanitarian AI paradox” of rapid individual adoption outpacing institutional readiness. 

📖 Understanding the Digital Divide and Humanitarian Principles, by A. Zwitter and K. Landicho (2024) highlights how AI offers transformative potential for early warning and disaster response in Southeast Asia, but warns that without addressing local digital divides, infrastructure gaps, and principled governance, these tools risk remaining out of reach or undermining humanitarian values. 

📖 The GBV AoR Helpdesk brief, J Ward, S. Spencer, K. Kalsi (2023) warns that unequal access to safe, inclusive AI tools in humanitarian response can deepen risks for women and girls, and calls for investments in local capacity, digital literacy, and gender-responsive design to bridge the AI access gap. 

📖 Innovation and Technology in Humanitarian Settings, UN Women (2021) UN Women’s Asia-Pacific good practice brochure showcases how inclusive innovation in humanitarian settings, such as mobile-based early warning systems and AI-powered aid delivery, can empower crisis-affected women, but only when paired with efforts to close gendered gaps in digital access, skills, and trust. 

Skill Up 

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

CLEAR Global’s Guide to Human-Centred Technology Design in Humanitarian Action (free) outlines practical, human‑centred design principles (including trust, two-way communication, digital skills, data protection, and inclusiveness) for participatory co‑creation of digital tools with crisis‑affected communities. 

NetHope’s Gender Equitable AI Toolkit (free) This toolkit encourages NGOs to consider gender equity as a fundamental aspect of AI program development. The toolkit is available as an e-learning course. 

Humanitarian Leadership Academy’s Women, Tech and Bridging the Digital Divide webinar (free) features frontline women technologists sharing real-world innovations and practical lessons in ensuring AI and digital tools reach underserved humanitarian communities 

UNESCO’s Red Teaming Playbook (free) democratizes access to AI safety testing by equipping humanitarian and civil society actors with practical tools to detect and address gender-based harms in generative AI. 

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

Humanitarian AI Today, hosted by Brent Phillips with Aral Surmeli - Founder & CEO of HERA Digital Health (finalist at the AI for Good Summit) 

In this Humanitarian AI Today podcast episode, physician and founder Aral Surmeli shares how HERA uses AI-powered chatbots to help refugee women and children access health services, discusses the challenges of digital divides and ethical AI use in humanitarian settings, and emphasizes the need for agile, context-aware tech built with and for affected communities. 

🕐 Run time:  ~60 min

Stay ahead of funding calls and events. 

💰 Humanitarian Grand Challenge – Transition-to-Scale Fund - Deadline: rolling basis  
Offers CAD $300K–$1.5M for locally led innovations that improve the lives of conflict-affected communities, with a focus on ethical design, and solutions for hard-to-reach populations. 
🔗 More info

🗓️ AI in the Humanitarian Sector – Research Report Launch (free, online) - When: 5 August 2025, 12:00 UTC. Hosted by Humanitarian Leadership Academy & Data Friendly Space, this online event unveils findings from a global survey of 2,500 humanitarian professionals on AI adoption, featuring practitioner panels and audience Q&A. 📎 More info

💰 Eureka Network – Disaster Resilience, Response & Recovery Projects - Deadline: 31 October 2025 
International R&D funding (range EUR €150K- €5M ) for collaborative projects (min. two organisations across Eureka member countries) developing AI, digital tech, materials, or devices to enhance disaster resilience, emergency response, and post-crisis recovery. 
🔗 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.