A framework for ethical AI at the United Nations
Catalogues a widely-cited UN ethical AI risk taxonomy - useful as a reference point for agencies developing their own risk registers.
Key points
- MIT AI Risk Repository summarises a UN-focused ethical AI framework identifying 13 AI risk categories.
- The framework covers risks relevant to APS governance work: bias, transparency, manipulation, and exclusion.
- The underlying paper is from 2021; this is a secondary summary with limited new analytical value for APS readers.
Implications for Australian agencies
- Monitor Risk and governance teams may want to monitor the MIT AI Risk Repository as a reference source when building or benchmarking agency AI risk registers.
Implications are AI-generated. Starting points, not advice — see methodology for how they're framed.
View original source
Copied.
"A framework for ethical AI at the United Nations"
Source: MIT AI Risk Repository – Blog
Published: 2 October 2024
URL: https://airisk.mit.edu/blog/a-framework-for-ethical-ai-at-the-united-nations-2
The MIT AI Risk Repository has published a summary of Lambert Hogenhout's 2021 paper proposing an ethical AI framework for the United Nations. The framework systematically lists 13 AI risk categories including incompetence, privacy loss, discrimination, bias, lack of transparency, deception, manipulation, and exclusion. It aims to align UN AI development and use with human ethical values and references practical tools such as assessment lists. This is a secondary summary of a three-year-old arXiv preprint rather than new primary research or policy.
Implications for Australian agencies:
- [Monitor] Risk and governance teams may want to monitor the MIT AI Risk Repository as a reference source when building or benchmarking agency AI risk registers.
Retrieved from SIMS, 18 July 2026.