Some UN agencies are exploring shared computing and data systems that could give them greater control of sensitive information and reduce reliance on US cloud and AI providers; officials are also considering exit plans for tech agreements, ITU CIO Khuloud Odeh told Semafor.
Odeh cited last year’s US sanctions on the International Criminal Court and Anthropic’s Mythos model cutoff earlier this year as “wake-up calls” about reliance on single, mostly American companies. She said this is not a wholesale move away from US technology: agencies are looking to create backup options and sort systems by sensitivity.
At a May meeting in Valencia, UN technology chiefs discussed sanctions-driven service suspensions and extraterritorial data access as risks to cloud arrangements; concerns also include possible US restrictions on chips and AI models. A UN registry lists more than 1,000 AI use cases, including projects to reduce hunger and connect schools. Odeh said agencies are shifting from experiments toward deciding where to invest, while UN data includes sensitive refugee records and health information.
