Algorithms, Artificial Intelligence and Discrimination

Automated systems now help decide who gets a loan, a job interview, welfare benefits or extra police attention. Because they learn from historical data, they can reproduce and scale up past discrimination, even when no one intends it. A 2018 study by researchers Joy Buolamwini and Timnit Gebru found that several commercial facial-analysis systems were far less accurate for darker-skinned women than for lighter-skinned men. Such findings raise human rights questions about equality, privacy, due process and the right to an explanation of decisions.

International responses have multiplied. UNESCO adopted a Recommendation on the Ethics of Artificial Intelligence in 2021. The EU's AI Act, adopted in 2024, sorts systems by risk and bans some uses. In 2024 the Council of Europe opened its Framework Convention on Artificial Intelligence, described as the first binding international treaty on the subject. Supporters of regulation stress protection; others warn that heavy rules could slow beneficial innovation. Most agree on the need for testing, transparency and human oversight.