What does AI policy design mean for humanitarian data collection and management?

WASP Power: An Ethical and Moral Crisis

The recent ICRC People Centric Humanitarian Response in Conflict conference September 2023 mirrored much of the thinking behind the outcomes of World Humanitarian Forum 2016, both highlighting the continued asymmetry of power between humanitarian organisations and the vulnerable populations they aim to assist. (Research papers over the years have consistently highlighted similar issues concerning the underlying problems of the Western aid system; see references at the end of this article). This power imbalance creates both a moral crisis and an ethical dilemma, now exacerbated by the role of generative artificial intelligence (GAI) and data analytics in humanitarian work.

While the advent of data analytics and artificial intelligence (AI) has significant potential to improve humanitarian efforts, a lack of accountability in data management poses a serious ethical concern. The Ada Lovelace Institute and the Berkman Klein Center for Internet & Society for example are doing fundamental work to forward discussions regarding the ethical implications of AI and data use, emphasising the need for greater accountability and equity in its use cases.

Data in Humanitarian Operations

RealtimeAid, a German NGO, has seen firsthand how easy it is for donors to be extractive when it comes to collecting personal information from vulnerable people, especially when it comes to compliance and managing risk. “We have to provide all sorts of personal information because there is a systemic lack of trust in local people who don’t speak English, if we don’t comply we won’t get our funding. The aid system is all wrong,” Tarek Alsaleh, Founder. Problematically, there is no real accountability for Western donors in terms of how they extract, or manage data from people who are living through war and crises. And this lack of accountability in aid is being exemplified by the advent of GAI. Therefore it matters who is behind the legal framework for AI policy - because that regulation will impact all users, especially the most vulnerable; such as children, groups stigmatised by society, people at risk. What happens then if that AI regulation is dominated by WASP-type thinkers?

The Demographics Behind AI Policy Design

One critical aspect exacerbating this issue is the lack of diversity in AI policy-making. This was evident in the Senate's "AI Insight Forum," where the majority of attendees were CEOs of tech companies, which is a conflict of interest situation given their vested interests - as well as being mainly WASP’s, thereby ensuring a discussion based on a specific set of shared values and life experiences. This homogeneity of people at the table will inevitably mean 'groupthink,' limiting the possibility and creativity of how policy can encompass diverse interest groups.

Humanitarianism and AI policy

The Humanitarian Aid ecosystem will be ever more reliant on AI. The vast capabilities of GAI will take-over many of the functions of the current humanitarian system, and will hopefully be the revolution this writer would like to see in the Western aid system. Whatever the future looks like for AI policy, it is certain humanitarians must also advocate to integrate humanitarianism into the very fabric of AI policy creation.

References

Ready to embark on your AI journey? Visit http://www.faid.io/ today and discover how FAID can support your NGO in harnessing the power of AI.

FrontlineAID is a social enterprise specialising in innovative processes and use of tech and artificial intelligence to support the localisation of aid. Frontline cuts away red tape so people can do good not paperwork.

Written by Ummul Choudhury

Thought leader, author, expertise on AI and localisation of humanitarian aid