“The AI discovered a mechanism that redressed initial wealth imbalance, sanctioned free riders, and successfully won the majority vote.”
–
DeepMind researchers, studying policy innovation through AI. In a new paper, published in
Nature Human Behaviour, the UK-based team reports on their “human-in-the-loop research pipeline called Democratic AI” that ran an online investment game with a thousand human players. “By optimizing for human preferences, Democratic AI offers a proof of concept for value-aligned policy innovation,” write the researchers.