Meta Study Finds Leading AI Models Avoid Criticizing Repressive Governments

Governments may be quietly shaping what AI will say about them, and this Meta study reveals a troubling bias you need to see.