How humanitarian actors can help local knowledge get found

Ask an AI tool about the humanitarian situation in the Democratic Republic of Congo (DRC), and you will get an answer in seconds. It may tell you about conflict, displacement, hunger, disease outbreaks and the millions of people who need humanitarian assistance.

Now look at where that information comes from.

When we asked ChatGPT, Gemini, DeepSeek and Perplexity humanitarian questions about the DRC, we later identified 100 verifiable citations behind their answers. None came from an organisation based in the country.

That finding stayed with me. The DRC has many people and institutions producing knowledge about its humanitarian situation. Congolese journalists report on the crisis, civil society organisations document what communities are experiencing, and national institutions and humanitarian organisations produce assessments and analysis.

The DRC was the most striking example of a wider pattern. Across 10 African countries, our African Humanitarian AI Citation Index (AHACI) examined 1,083 verifiable citations used by four major AI platforms. Only 5.5% came from organisations based in the country being discussed, while 68.8% came from the UN or other multilateral organisations.

The study recorded the sources visibly cited in the AI-generated answers. Questions about why individual sources appeared, whether suitable local evidence existed for every query and how source quality compared across organisations fell outside its scope.

The findings still raised a practical issue that reaches beyond AI: local organisations can produce valuable humanitarian knowledge, publish it online and remain difficult to find.

Localisation is also about whose knowledge gets seen

The humanitarian sector has spent years debating localisation, including who receives funding, who leads responses and who has meaningful influence over decisions.

The debate remains urgent. ODI Global’s 2026 analysis of humanitarian financing found that just under 10% of international humanitarian funding reached local and national actors directly or indirectly in 2024, well below the Grand Bargain’s 25% ambition.

Knowledge is part of the same power relationship. A previous HPN article on localising knowledge generation described how local organisations can be heavily involved in gathering information, while other institutions retain greater influence over analysis, publication and dissemination. The experience showed why local actors need meaningful participation throughout the knowledge-generation process.

That conversation can continue after the research has been completed. Local evidence loses some of its reach when people struggle to discover it, identify its producer or trace it back to the original source.

A report can be online and still be hard to find

Think about what happens after a humanitarian assessment is completed. Teams gather information from communities, other staff analyse it, and a report is written, reviewed and uploaded to a website. Six months later, someone unfamiliar with that organisation may struggle to find the report. It may sit several clicks deep inside a website under a title such as Rapid Assessment Report No. 4. The geographical area may be unclear, the publication date buried inside the document, and the webpage hosting it may contain little more than a download button.

These features make information harder for people and digital systems to identify, understand and retrieve. They also show why putting a document online is only one part of making knowledge accessible.

The humanitarian sector already works with this principle in data management. OCHA’s Humanitarian Data Exchange (HDX) asks contributors to provide metadata that helps users understand and use datasets. Its guidance also encourages meaningful titles because the words used in a title affect how easily people can find relevant information.

The same lesson applies more widely to humanitarian publishing: how information is published affects how easily it can be found and used.

What happens after you press publish?

When it comes to knowledge-production, humanitarian organisations understandably invest most of their effort in collecting evidence, analysing it, writing a report and getting it published.

Publication starts another part of the journey. The information still needs to be discovered, understood, attributed and used by people who may have never encountered the organisation that produced it.

Search engines have shaped this process for years. Generative AI has added another route. A journalist, researcher or policymaker may begin with an AI-generated overview and then follow the sources behind the answer. This makes how humanitarian evidence is presented, attributed and preserved increasingly important. There are several practical ways organisations can improve that journey.

1. Give important evidence a proper digital home

For important assessments, research and evaluations, a downloadable PDF can sit alongside a simple webpage explaining what the publication covers, where and when the information was collected, who produced it, the main findings and where to access the full report.

This requires little technical sophistication. It gives the evidence context and creates a clear place for people to find, understand and share it.

Someone encountering an important report for the first time should be able to understand what it contains, where it applies and why it matters.

2. Name reports for people who do not already know you

A title such as Food insecurity and displacement in northern Mozambique: rapid assessment, July 2026 tells a reader far more than Rapid Assessment Report 07/26. It immediately identifies the issue, geography, report type and period.

HDX recommends the same basic discipline for humanitarian datasets, including avoiding unclear abbreviations and using words that help people identify relevant material.

Clear titles reduce the prior knowledge someone needs to decide whether a publication is useful. Humanitarian knowledge becomes easier to discover when its meaning is clear outside an organisation’s internal filing system.

3. Make attribution travel with the evidence

Local evidence often passes through several organisations.

A national organisation may collect and analyse information that later feeds into a wider international report. The international publication may receive greater circulation, and over time the original source can become less visible. This creates both discoverability and attribution issues.

Local organisations can strengthen attribution by clearly stating their full organisational name, publication date and geographical coverage, and by providing a suggested citation for substantial reports.

International actors also play an important role. When local evidence informs a wider publication, they can link to the original source where appropriate, name the organisation clearly and preserve attribution when findings are reused.

Localisation is partly about recognition and power. The organisations producing knowledge should remain visible as that knowledge moves through the humanitarian system.

4. Make evidence easier to understand and reuse

Humanitarian reports are often long because humanitarian situations are complex. Readers still need a clear entry point.

A useful publication should quickly explain what happened, where it happened, who is affected, what the evidence shows and how the information was collected. Important statistics need context, charts need clear labels, and methodology should be easy to locate.

Where data can safely be shared, organisations can also provide it in formats that allow reuse instead of leaving important information trapped inside a PDF. The humanitarian sector already has useful experience here. The Humanitarian Exchange Language (HXL) was developed to make humanitarian datasets easier to exchange and combine.

The principle extends beyond spreadsheets. Clearer information serves people first and also gives digital systems a better chance of interpreting the material accurately.

5. Keep useful evidence alive

Knowledge can become harder to find long after publication.

A website redesign may change URLs, move old documents or break links from other organisations. For a team responding to today’s emergency, maintaining a report published several years ago may feel less urgent than producing today’s update.

Humanitarian crises have histories. Researchers, journalists, communities and humanitarian actors often need earlier evidence to understand how conditions have changed. Stable links and usable archives are therefore part of humanitarian knowledge stewardship. They help preserve evidence long after the immediate communication around a report has ended.

6. Make visibility responsible

Greater visibility also requires careful judgement.

Humanitarian organisations hold information that can expose vulnerable people and communities to additional risk. The IASC Operational Guidance on Data Responsibility emphasises the safe, ethical and effective management of humanitarian data.

The same principle applies to decisions about discoverability. Organisations need to consider what information should be publicly accessible, in what form and with what safeguards. In some cases, a summary or metadata may be appropriate, with access to the underlying information remaining controlled.

Better discoverability and protection therefore need to develop together.

Language shapes what gets seen

All AHACI questions were asked in English, which is an important limitation.

Several countries in the study are Francophone or Lusophone, and humanitarian knowledge across Africa is produced in many other languages.

The extent to which language affected the citation patterns remains an area for further research. Valuable evidence published mainly in French, Portuguese or another language may travel differently when users search in English.

A practical response is to consider short summaries of important reports in additional relevant languages where resources allow. This can help knowledge move beyond its original context without expecting local organisations to reproduce every publication in multiple languages.

Knowledge localisation therefore needs to consider both who produces evidence and how that evidence moves across languages.

Visibility should be a shared responsibility

Many national and community organisations operate with small teams, limited communications budgets and fragile digital infrastructure. ODI Global’s work on local organisations’ overheads shows how humanitarian financing can leave local actors with limited support for the organisational capabilities they need to remain sustainable.

International organisations often operate with stronger digital infrastructure, dedicated information-management teams, translation capacity and global distribution networks. That makes knowledge visibility a shared responsibility.

When local partners produce useful evidence, international actors can link directly to it and preserve attribution. Donors can recognise digital publishing, knowledge management and archives as legitimate organisational capabilities. Organisations with greater reach can use it to amplify local knowledge and direct audiences to the original source.

The wider challenge is creating a humanitarian information system in which local expertise has a fair chance of being discovered and recognised.

Localisation should include knowledge visibility

Localisation has rightly pushed the humanitarian system to move funding, leadership and decision-making closer to crisis-affected communities. Knowledge belongs in that conversation.

AI citation is one signal of a wider shift in how people discover information. A more useful measure is whether someone unfamiliar with a local organisation can find its evidence, identify who produced it and trace important claims back to the original source.

Improving that journey may involve ordinary changes: a clearer title, a useful webpage, a stable link, better attribution or a concise summary alongside a longer report.

These steps address one part of the wider inequalities that shape whose knowledge carries authority in the humanitarian system. They can make it easier for valuable local evidence to retain its identity, travel further and remain accessible after publication.

As AI and search become more common routes to humanitarian information, localisation may need to make room for one more question alongside who gets funded, who gets heard and who gets to decide: Who gets found?

Hezron Ochiel is an award-winning strategic communications and public relations professional, researcher and founder of Hezron Insights. His work examines how artificial intelligence, search and digital publishing are changing visibility, institutional authority and access to information in Africa.