This week saw major steps in AI safety rules. The World Digital Technology Academy launched new safety tests for single AI agents at a UN meeting. These tests act like seatbelts for AI helpers in cars, hospitals, and banks.

In America, Texas passed a big law requiring AI systems to be clear and fair. Michigan proposed a new bill making AI companies share safety plans and protect whistleblowers. California is debating laws about human control over AI in important systems like power grids.

Doctors got warned to get patient permission before using AI for health decisions. The UN held a special meeting where experts said we must understand AI risks before putting them everywhere. BRICS countries asked the United Nations to create fair worldwide AI rules.