Thomson Reuters Invests $40M to Accelerate In-House Legal and Tax AI, Moving Away from External Dependencies

Release Overview

Thomson Reuters has announced a $40 million investment to strengthen its proprietary AI capabilities, aiming to break away from its reliance on external AI technologies. This marks a strategic shift from the model of "borrowing" existing general-purpose AI models provided by companies like OpenAI and Anthropic, toward owning and controlling its core technology to specialize in professional domains.

Purpose of the Investment

This investment will be allocated to building and optimizing proprietary AI models by leveraging the vast specialized data assets in legal, tax, accounting, and media that Thomson Reuters has accumulated over many years. In professional tasks where reliability is paramount, the company aims to establish a foundation that generates high-precision answers difficult to achieve with general-purpose models.

Technical Background and Strategy

While many companies utilize third-party AI models via APIs, maintaining data privacy and continuously improving model accuracy remain major challenges. By training its proprietary AI engine on top-tier professional documents, Thomson Reuters plans to gradually reduce its dependence on external providers and establish a unique competitive advantage.

Future Outlook

The company has determined that owning proprietary AI will become the most powerful source of competitive advantage in industry-specific generative AI moving forward. Moving ahead, Thomson Reuters is poised to accelerate service deployment centered around its dedicated AI infrastructure, further enhancing solutions tailored for professionals.