Humanitarian AI: The hype, the hope and the future

Killer robots and autonomous weapons. Global unemployment. Mass surveillance.
Superintelligence. The Singularity. There are a host of fears about artificial intelligence (AI) and related industries such as machine learning (ML). But what exactly are they? How are they being used? What, if anything, do they offer the humanitarian aid industry? Are we at the beginning of a fundamental transformation of a business model that, in many ways, is no longer fit for purpose? Will they trigger an industry-wide digital disruption that will challenge the dominance of the largest humanitarian aid organisations, or serve to exacerbate the North–South divide?

Humanitarian actors and their donors are only just beginning to explore the ways in which these technologies will impact humanitarian action. Where practical, the term ‘humanitarian action’ is used in place of humanitarian aid to expand analysis and discussion beyond the delivery of services and support and include work linked to identifying and understanding conflict drivers and dynamics. See Glossary. This Network Paper attempts to explore the benefits, opportunities, risks and obstacles to using AI/ML in the humanitarian sector. It seeks to unpack the myths and rhetoric related to AI/ML and evaluate the range of arguments made in favour of or against their use, drawing on literature and interviews with scores of experts across the aid and technology industries. Lastly, the paper offers some conclusions and suggestions for how humanitarian actors, technologists and donors might engage with AI/ML in humanitarian contexts.

The paper is in five sections:

• Chapter 1: Summary. A review of the paper’s key findings, some emerging lessons and suggestions for future engagement.
• Chapter 2: ‘Wake up and smell the technology’: the opportunities and benefits of AI/ML. An overview of the opportunities and benefits that AI/ML offers the humanitarian sector.
• Chapter 3: The harm and the hype: the risks of deploying AI/ML for humanitarian action. An exploration of the risks and limitations of using AI/ML in the humanitarian sector.
• Chapter 4: The runway to scale: enduring obstacles to deploying AI/ML. An analysis of the factors that limit and enable the widescale use of AI/ML for humanitarian action.
• Conclusions: What next for humanitarian AI? Emerging conclusions on how humanitarian actors might think through ways to engage with these technologies. Hyperlinks to additional resources have been provided throughout the paper.

Hyperlinks to additional resources have been provided throughout the paper.