This week's AI news shows that building trust between humans and AI agents is becoming more important than ever. Companies are working hard to make sure AI agents do exactly what people ask them to do and nothing more.

Mastercard introduced something called Verifiable Intent, which works with Google to create a trust system for when AI agents shop for you. This new system keeps a record that proves what you told the AI to do. It's like having a receipt that shows the AI followed your instructions correctly.

However, there are still big problems with AI agents. Security experts found that AI agents can make mistakes and get confused. Hackers can trick them with special commands. Companies say they need people to watch over AI agents to make sure they don't break things or cost too much money.

Trust became a huge issue this week when ChatGPT had problems that made many people delete it from their phones. People realized they need to be careful about which AI systems they trust with their personal information.

The good news is that a large group of people including workers, teachers, and leaders created something called the Pro-Human AI Declaration. This plan says AI systems should always stay under human control and should protect kids and communities.

Companies are also discovering that 70% of businesses now use AI agents, but many worry about security problems. The lesson this week is clear: AI agents are becoming common, but people must be able to trust and control them before they become part of our everyday lives.