OpenAI rolled out Astra for Law on Thursday, a specialized version of its GPT-6 Astra technology built specifically for law firms and legal software providers. The platform combines OpenAI's Astra model with an indexed database of US case law, statutes, regulations and other legal materials, along with specialized instructions for legal analysis and writing. It's the latest salvo in an intensifying competition among major AI labs to capture market share in the legal services sector.
The race isn't new. Google expanded its Gemini Enterprise offerings for legal professionals last month, while Anthropic has offered tools for lawyers using its Claude AI assistant since January. Thomson Reuters, which owns the Westlaw legal research database, launched Thomson 1.0 last month—a model trained on its own authoritative legal content and built on open-source technology. The market is crowded, and the stakes are high.
The Competitive Landscape
OpenAI's entry into legal AI reflects a broader strategy: let private companies build on top of your technology rather than trying to dominate every vertical yourself. The company said that legal AI vendors including Harvey and Legora will be able to build products on Astra for Law. The platform will also integrate with major software providers like Relativity, Clio, Intapp and Thomson Reuters—a move that positions OpenAI as infrastructure rather than a direct competitor to existing legal tech firms.
This approach differs from Thomson Reuters's strategy. The company developed Thomson 1.0 using its own proprietary legal content accumulated over decades, creating a closed ecosystem where its training data becomes a competitive moat. Thomson Reuters controls the data; OpenAI is betting it can win through superior model architecture and open partnerships.
Market Access and Risk Management
OpenAI said "selected" law firms will gain initial access to Astra for Law through a special program that includes protections for confidential client work. That's a critical detail. Law firms handle sensitive information, and any AI platform operating in this space must address data security and attorney-client privilege concerns. The company tested its product with major firms including Sullivan & Cromwell, Ropes & Gray, Cooley, Latham & Watkins and Wachtell Lipton before launch.
The confidentiality protections suggest OpenAI understands the regulatory and reputational risks. A data breach or inadvertent disclosure of privileged information could trigger massive liability. By limiting initial access to selected firms and building in safeguards, the company is managing downside risk while gathering real-world performance data.
Why This Matters:
The legal AI market represents significant revenue potential, but it also illustrates how private enterprise competition drives innovation faster than government could mandate it. Three separate companies—OpenAI, Google, and Anthropic—are racing to solve the same problem, each with different approaches and business models. That competition benefits law firms and their clients through better tools and lower prices. The market is also self-regulating: firms won't adopt AI systems that don't protect client confidentiality, creating natural incentives for responsible development. OpenAI's decision to work with established law firms on testing, rather than imposing solutions top-down, reflects market discipline. The legal profession's gatekeeping role—bar associations, ethical rules, liability concerns—will ultimately determine which AI platforms succeed, not government mandates. This is how innovation should work: companies compete, customers choose, and institutions enforce accountability.