AI safety needs social scientists

OpenAI Blog · 7 years ago

Researchers at OpenAI argue that AI safety work must include social scientists to address how alignment algorithms function with real people rather than idealized humans. OpenAI plans to hire social scientists full-time to collaborate with machine learning researchers on this problem. This shifts AI safety research beyond technical engineering toward understanding human psychology, rationality, and decision-making biases that affect whether AI systems actually align with human values in practice.