Article by Paula Dupraz-Dobias: “Scrambling for solutions to reduce spending at the United Nations, a plan presented in late July – part of the UN80 reform initiative – referred to the use of artificial intelligence (AI) to cut duplication in reporting across the body’s numerous divisions without providing much detail as to how this may be achieved.
Tech tools involving data sourcing are increasingly being used to improve efficiency in humanitarian response. But partnering with private technology firms may pose risks for aid organisations as they increase their digital engagement.
Six years after the World Food Programme (WFP) announced an agreement with United States tech and data analysis firm Palantir to help streamline its logistics management, sparking a barrage of concerns over data protection, questions about humanitarians working with private companies have again resurfaced.
On Thursday, Amnesty International condemned the use by the US government of tools developed by Palantir and other tech firms to monitor non-citizens at pro-Palestinian demonstrations as well as migrants…Experts have recognised that more needs to be done to define the parameters of partnerships between aid organisations and the private sector to mitigate risks.
“The challenge today is how do you improve the way you make decisions or design services or develop policies that leverage new tools such as data and AI in a systematic, sustainable and responsible way … where no one is left behind,” says Stefaan Verhulst, co-founder of the The GovLab, a research centre at New York University focused on how to improve decision-making and leveraging data in the public interest.
Organisations, he explains, need to develop a more comprehensive data governance framework. That includes a clearer articulation of the reasons for collecting data and using AI as well as governance principles for the use of data based on human rights.
Local stakeholders, such as potential aid beneficiaries, should also be included in the decision-making, he says, while data should be managed using a life-cycle approach – from collection to processing to analysis and use. Sourcing data from vulnerable populations required a “social license” or contract, to use and reuse data, which Verhulst recognised was one of the key missing elements currently…(More)”
