ChatGPT says it won’t write offensive content. We set out to test whether that’s true.

Artificial intelligence can significantly impact human rights—both positively and negatively. Human rights impact assessments conceived specifically for AI are needed to prevent potential harms and reap ...

An infinite supply of plausible opinions from fake, AI-powered pundits threatens to crowd out genuine discourse.

The EU needs the technical standards supporting its AI Act to be restrictive enough to protect consumers, but flexible enough to enable innovation. Given society’s current understanding of AI, there are ...

For the first time, major world democracies have gone public with a set of common protections that they apply when accessing individuals’ personal data for intelligence or law enforcement purposes.