Thomson Reuters builds a Claude alternative on Chinese Qwen

What happened

Thomson Reuters is moving some of its legal AI tasks from Anthropic's Claude to Thomson-1, an in-house model built on Chinese open-source technology. The company expects this change to reduce costs, but it also raises questions about the security of using Chinese AI solutions.

Confirmed

Global impact / market context

This shift could lower Thomson Reuters' spending on outside AI services, which means more cash available for other business areas. However, relying on Chinese technology might create regulatory risks, potentially affecting their legal product reliability and customer trust.

Analyst inference

Companies are increasingly looking to replace expensive outside AI tools with cheaper in-house options to cut costs. Using Chinese open-source models could become more common, but security concerns may make investors cautious about companies that depend on such technology.

Analyst inference

What to watch

1. Watch whether Thomson Reuters completes its transfer of legal AI tasks to Thomson-1 and if it expands this technology to other parts of its business, which would confirm cost reductions. Confirmed
2. Investors might assess how the potential security risks from Chinese AI technology could impact Thomson Reuters's reputation with legal clients, who handle sensitive information and might demand stronger guarantees. Proposed
3. Watch if other large firms copy this move to adopt cheaper in-house AI, which could increase competition in the AI market and push outside providers like Anthropic to lower prices. Analyst inference

Evidence

• cryptopolitan.com