1 of 6

Building a

Humanitarian AI Code of Conduct Together

2 of 6

How did we get here

Lessons learned from biometrics & emerging tech
NetHope Global Summit workshop in Oct 2023
NetHope AI Working Group to take forward
Virtual follow-up workshop in Feb 2024
Ethics & Governance sub-group action as deliverable

Kicked off in Oct 2023 at the NetHope Global Summit with an in-person workshop. Followed up in Feb 2024 with a virtual workshop
This idea behind this is that AI technologies are on the rise – specifically genAI, the use of LLMs and NLP – and this is creating new opportunities as well as risk for humanitarian and development organisations.
I had personally come from this after spending a few years working on biometrics.

Biometrics have become ubiquitous over the last 20+ years, before the responsible data movement, before GDPR, before data privacy really had its spotlight. There are narratives of efficiency and fraud etc around biometrics, but they are a very sensitive kind of personal data and as a sector we have been witnessing the harm that has come from that data not being handled properly or outright being misused.
My work on this, introducing organisational policy and trying to do some advocacy around improving responsible use of biometrics, in recognition that it’s unlikely we can altogether stop the use of biometrics.
The learning from this is: let’s come together as a sector to think about these big questions, these big risks. Much of the risks and ethics around AI are well documented now, so we’re not talking so much about apocalyptic scenarios but rather, how can AI make inequality worse, influence decisions that impact people negatively, be used to repress civil society, and so on.
So taking the principle of Do No Harm even further. Applying AI to humanitarian contexts, getting on the same page, and agreeing to some really fundamental principles.

The AI Working Group in NetHope has taken this on as a deliverable. I’m a co-lead for that group, now leading the sub-group on Ethics & Governance.
So we’re aiming to have this CoC as a framework – nothing too intense and designed in a way that incorporates the leading concerns for member organisations, but also in an agile way, that recognises the pace of AI development, so we can adapt to those changes more quickly. We can build in that longer term resilience.
This is currently a NetHope project, with NetHope member organisations. Did think about expanding consultation out, but ultimately decided to limit the scope in the earlier iterations of this, so that we could focus on what Member orgs are thinking about, but are also doing some light touch external consultation as well.

So, this does not wholly represent civil society by any means, but we are aiming to have this as a tool to guide other related deliverables that are practical, so the CoC would serve as a foundation of agreement and then we’re going to work on building out a practical set of governance templates, so e.g. organisational policy, guidance, red lines, etc.

3 of 6

Code of Conduct (DRAFT)

Our use of AI technology and related data practices must:

1. Do No Harm and align with existing humanitarian principles and frameworks.

2. Have a net positive impact on our organisations’ missions and for the individuals and communities we serve, and not exacerbate the problems we are working to solve* .

3. Be fair, inclusive, accessible, and feminist; mitigate bias; and ensure transparency and explainability.

4 of 6

Code of Conduct (DRAFT)

Where there are sector-specific high risk concerns, we agree:

4. Not to use AI to generate photo-realistic images of children or programme participants for the purposes of publication, including campaigning and fundraising*.

5. A human will review and, where necessary, contextualise and attribute content generated by or with the help of AI for the purposes of publication.

6. A human will review any decision made or facilitated by AI that may cause real or perceived harm to an individual or community*.

7. To prioritise safeguarding and child protection and establish any additional guardrails needed where AI is used in these contexts*.

5 of 6

Code of Conduct (DRAFT)

We will work together to:

8. Collaborate and exchange knowledge and best practices, and collectively support digital literacy and skills in AI across the sector.

9. Align on the approach to due diligence, assessment, and procurement of AI vendors and tools, as well as limits on the commercial use of data.

10. Align where we have different regulatory environments and uphold the highest available standards and practice of data protection, privacy, and security.

11. Understand and address how informed consent manifests in AI contexts and whether it can be meaningfully used as a basis for collecting personal data.

12. Ensure effective governance, accountability, and pathways for redress.

13. Collectively engage with questions and address new risks as the AI landscape evolves*.

6 of 6

Next steps