Companies around the world are learning that trust is the most important part of using AI agents that work on their own. This week, major tech leaders came together to focus on how to build AI systems that people can really believe in.

One big announcement was from Thomson Reuters in Canada, who brought together companies like Anthropic, AWS, Google Cloud, and OpenAI to form the Trust in AI Alliance on January 14, 2026. This group is working to figure out what makes AI agents trustworthy, especially when they make important decisions in areas like law and finance. The leaders say that trust comes from building safety and accountability right into the computer code that makes AI work.

But there are big problems to solve. Experts are worried about fake AI agents that bad actors could use to trick organizations and steal information. When companies cannot tell if an AI agent is real or fake, they cannot safely let it access their most sensitive data.

Companies are also learning to build guardrails into their AI agents—kind of like safety rules that keep AI from doing things it should not do. Shopify and Salesforce showed that AI agents with strong protections are being used in real businesses to handle important work automatically. However, most leaders agree that AI should work alongside people, not replace them.

The big takeaway this week is clear: AI agents will only work if people can trust them. Companies that build trust into their systems from the start will succeed, while those that do not will struggle.