OpenAI introduces law-specific AI tool, Astra for Law

Currently available in the United States, with the expectation of rolling out to more countries in the near future, the new operation combines GPT-6 Astra, which is the company’s most recent and significant model, with settings, tools, and context specifically created for legal work.

OpenAI noted that it would focus on privacy and governance, giving law firms specific controls for client work to remain confidential.

What are the new changes?

The largest implementation is a legal-specific model configuration, with the new tool tailored with legal analysis and documentation rather than being a general-purpose model, like the other free regular platforms.

Furthermore, Astra for Law provides a dedicated legal search, allowing users to search a legal index that covers US case law, statutes, regulations, court rules, and administrative decisions, updated on a daily basis.

Astra for Law also creates stronger legal reasoning workflows, in the sense that it is designed to analyse client facts, develop arguments, and identify strong counterarguments.

VIEW ALL

Another important benefit outlined in the new tool is its firm-specific AI workflows and customisation.

This means that law firms can build Astra for Law around their own precedents, methods, standards, and judgment.

OpenAI has already developed custom applications, with firms such as Sullivan & Cromwell, Ropes & Gray, Cooley, and Skadden, including agreement analysis, deal diligence, IPO preparation, and regulatory-risk workflows.

AI as a complement to legal expertise

Speaking to Lawyers Weekly, Maddocks director of operational excellence Mitch Grant highlighted how AI can accelerate research while retaining the need for human judgement.

“Tools like Astra for Law can improve the speed and efficiency of legal research; they can surface relevant cases and legislation faster than traditional methods,” Grant said.

“However, they will not replace the need for human judgement; lawyers must still interpret the law in context and exercise strategic thinking.

“They must assess the strength of arguments and advise clients on risk.”

Ultimately, Grant framed AI as a tool that enhances rather than replaces legal expertise.

“AI tools are best understood as a complement to professional expertise rather than a substitute for it.

“The most important part of the process remains the ability of a lawyer to apply judgement to the facts,” he said.

Privacy, governance, and responsible AI implementation

Grant pointed out that privacy is a key consideration for firms that are handling sensitive client information through AI platforms.

“Law firms should treat privacy as a serious concern when adopting any AI platform,” Grant said.

“These tools process large volumes of sensitive client data; that data may include privileged communications and confidential case materials.”

Grant recognised that firms need to assess AI tools against their legal and professional requirements before looking to adopt such technologies.

“Firms must ensure that the platform complies with the Australian Privacy Act 1988 and any applicable professional conduct obligations.

“Firms should assess the data-handling practices of any AI provider before adoption,” he said.

Similarly, Maddocks CEO Paul Woods identified the importance of governance and safeguards when interacting with AI in the legal profession.

“The value of AI in legal practice depends on the governance framework around it,” Woods said.

“Firms that adopt AI tools with proper safeguards and clear policies will benefit from efficiency gains without compromising client trust.

“They should also invest in training to ensure that lawyers understand the limitations of AI-generated outputs.

“The firms that approach this thoughtfully will be well-positioned to use these tools effectively while maintaining the standards their clients expect.”