Thomson Reuters built its own AI model off Chinese tech to rely less on Claude (www.businessinsider.com)

Thomson Reuters has announced the launch of its AI model, Thomson-1, which is designed to reduce its reliance on expensive AI solutions from Anthropic, notably Claude. This new model is based on Snowdon, a modified version of Alibaba's open-source Qwen framework. By constructing an in-house AI model, Thomson Reuters aims to leverage its expertise, starting with applications in document review while still maintaining partnerships with companies like Anthropic for other functionalities. This development is significant for the AI/ML community as it reflects a growing trend among large tech firms to build proprietary models using open-source technologies, particularly from China, which may offer cost-effective alternatives to established systems. However, this shift also raises concerns about security and ethical implications, especially in light of allegations against Chinese labs regarding the unauthorized use of AI outputs. Thomson Reuters is addressing these issues by ensuring its model is ethically developed and free from bias. The move toward in-house AI represents a strategic pivot that emphasizes intellectual property and long-term value over continuous dependency on external sources.