A framework for ethical AI at the United Nations

2 Oct 2024 · MIT AI Risk Repository – Blog Global

A catalogued AI risk taxonomy from a UN context offers comparative value for agencies building or auditing their own risk registers - though more current Australian frameworks exist.

Key points

Summary

The MIT AI Risk Repository has summarised Lambert Hogenhout's 2021 paper proposing an ethical AI framework for the United Nations. The framework enumerates 13 AI risk categories - ranging from bias and lack of transparency through to lethal autonomous weapons, manipulation, and exclusion - and argues for practical assessment tools to ensure AI development aligns with human ethical values. The paper explores multiple ethical frameworks and proposes a structured path for UN adoption. As a 2021 arXiv preprint contextualised for UN institutions, it predates more recent Australian and international risk frameworks but may still inform comparative taxonomy work.

Implications for Australian agencies

Implications are AI-generated. Starting points, not advice.