OpenAI Astra for Law Brings GPT-6 AI to Legal Work

OpenAI is expanding its Artificial Intelligence technology into the legal industry with Astra for Law, a version of its latest model, GPT-6 Astra, tailored specifically for professional legal work.

Announced on Thursday, September 17, the new offering is designed to help lawyers with tasks including legal research, document analysis, drafting and transaction review. OpenAI said the system has settings, tools and contextual capabilities adapted for the way professional legal teams work.

The initial rollout will be limited to a small number of law firms through Trusted Access in ChatGPT and Codex. Rather than positioning the technology as a replacement for lawyers, OpenAI says Astra for Law is intended to work alongside a firm’s existing expertise, processes and professional judgment.

The move places OpenAI deeper into a legal technology market where artificial intelligence companies are increasingly competing to become part of established professional workflows. For law firms, the attraction extends beyond conversational AI to the ability to connect models with large collections of documents, firm knowledge, specialist databases and existing legal technology systems.

What Is OpenAI’s Astra for Law?

Astra for Law is designed around the specific requirements of legal professionals rather than being presented simply as a general-purpose AI chatbot.

OpenAI says the system can assist with a broad range of legal activities, including researching matters, analysing documents, drafting content and reviewing deals. Its legal industry examples also cover areas such as initial public offerings, document review, deal diligence and working with a firm’s own precedents and playbooks.

This focus is important because legal work often involves large volumes of information that must be reviewed and organised before a lawyer can make a professional judgment. Contracts, case files, previous agreements, internal precedents and other documents can contain thousands of pages of potentially relevant information.

AI systems can help process that material, identify information and assist with drafting. The lawyer remains responsible for evaluating the output and applying professional judgment to the matter.

OpenAI Wants AI Embedded in Existing Legal Workflows

One of the most notable aspects of the announcement is OpenAI’s emphasis on integration rather than standalone chatbot use.

The company is seeking to connect its models with technology that law firms already use. Partnerships with legal technology providers including Clio, DeepJudge, iManage, Intapp, Relativity and Thomson Reuters are intended to make it possible to bring firm knowledge, case information, documents and established workflows into ChatGPT.

These integrations are designed to preserve existing permissions and controls while allowing AI models to work with information already maintained within a firm’s technology environment.

For a law firm, that distinction can be significant. A general AI assistant may be able to generate a draft or explain a legal concept, but a professional legal system needs access to the firm’s own documents, precedents, databases and processes to become useful across day-to-day work.

OpenAI’s approach therefore appears focused on making its models part of the Infrastructure lawyers already depend on rather than asking firms to create an entirely separate workflow around an AI application.

Legal Research and Document Analysis

Research and document analysis are among the areas where Astra for Law is expected to provide practical assistance.

Legal professionals frequently need to examine large collections of documents, identify relevant information, compare provisions and organise findings before drafting advice or preparing for negotiations. AI can potentially reduce the amount of manual processing involved in these stages.

OpenAI’s legal offering is designed to assist with these activities while keeping lawyers involved in the process. The system can help analyse material and prepare work product, but legal professionals remain responsible for assessing whether the information is accurate, relevant and appropriate for the specific matter.

This human role is particularly important in legal work because an apparently small factual or interpretive error can have significant consequences for a client, transaction or case.

Drafting and Deal Review Are Key Use Cases

OpenAI is also targeting drafting and transaction-related work. Its examples include deal diligence, document review and the use of a firm’s own precedents and playbooks.

In transactional practices, lawyers can spend substantial amounts of time comparing agreements, identifying provisions, checking documents against established standards and preparing drafts. An AI system connected to relevant firm knowledge could assist with these repetitive information-processing tasks.

For example, a legal team could use an AI system to help organise information from a collection of transaction documents before lawyers conduct a more detailed review. Similarly, firm-specific precedents and playbooks can provide context for drafting and review.

The intended workflow is therefore not simply “ask AI a question and accept the answer.” Instead, the technology is being positioned as an assistant that can process information and prepare material within professional workflows where lawyers continue to exercise judgment.

Confidentiality and Data Protection Take a Central Role

Data protection is particularly important when AI is introduced into legal environments because law firms handle confidential client information, privileged communications, sensitive transaction documents and litigation material.

OpenAI said its legal products are designed around professional obligations such as protecting client confidentiality and maintaining independent professional judgment.

The company’s legal offering also includes specific data-handling arrangements. According to OpenAI, its API usage comes with zero retention, while ChatGPT Enterprise uses an “eyes-off” human review posture by default.

These controls are intended to address concerns that can arise when firms consider placing sensitive legal information into AI systems.

Data governance, however, remains an important part of any firm’s implementation decision. Law firms must consider their own confidentiality obligations, internal security policies, client requirements and regulatory responsibilities when determining how AI tools should be used.

Why Law Firms Are Becoming an Important AI Market

The legal sector presents an attractive market for AI companies because much of its work involves information-intensive and repetitive knowledge tasks.

Lawyers may need to review extensive document collections, search for specific information, compare contracts, research legal issues and prepare drafts. These activities require expertise, but many also involve substantial amounts of information processing that AI systems can potentially assist with.

That combination has encouraged major AI companies to develop products specifically for professional legal users.

For OpenAI, the opportunity extends beyond providing lawyers with access to a general-purpose model. The company is attempting to combine AI reasoning with legal databases, firm-specific knowledge, established workflows and specialist software.

This approach could make AI more deeply integrated into how legal teams work while preserving human responsibility for decisions that require professional judgment.

Anthropic Is Also Expanding Into Legal AI

OpenAI’s legal push comes as other major AI companies increase their focus on professional services.

Anthropic has also expanded Claude’s presence in the legal sector with tools and integrations aimed at lawyers and law firms. Earlier this year, the company introduced access to 12 legal practice plug-ins covering areas including litigation, employment and commercial law.

Anthropic said these tools could be deployed through its Cowork platform or embedded into law-firm systems.

The competing approaches reflect a broader shift in the AI market. Instead of competing only over general-purpose chatbots, AI companies are increasingly developing specialised products for industries where organisations have large amounts of structured and unstructured information.

Legal services are particularly suited to this strategy because firms already maintain extensive collections of documents, precedents, research material and professional workflows.

The Difference Between a Legal AI Assistant and a Lawyer

The expansion of AI in legal services also highlights an important distinction between assisting with legal work and replacing professional legal judgment.

A model can potentially help identify information, summarise documents, organise research and produce a draft. But the output still needs to be assessed within the context of the specific client, matter and applicable legal requirements.

Lawyers must also determine whether information is reliable, whether an argument is legally sound and whether a document appropriately reflects the client’s interests and instructions.

OpenAI’s positioning of Astra for Law reflects this distinction. The company says the system is designed to work alongside a firm’s expertise and methods rather than replace them entirely.

That approach also recognises that legal work involves responsibilities that cannot simply be delegated to an automated system. Independent professional judgment and responsibility for client matters remain with legal professionals.

Firm-Specific Knowledge Could Be a Major Advantage

One of the potentially important features of OpenAI’s approach is the use of a firm’s own precedents and playbooks.

Every law firm can develop its own approaches to drafting, transaction review, research and internal processes. A general-purpose AI model may know broad legal concepts, but a system connected to a firm’s own approved material can potentially provide assistance that is more closely aligned with the firm’s established practices.

Integrating that knowledge also changes how AI is used. Instead of asking a model to work only from a user’s prompt, lawyers can potentially work with a system that has access to authorised information within the firm’s existing technology environment.

OpenAI’s partnerships with legal technology companies are aimed at enabling that type of workflow while retaining existing access permissions and controls.

AI Could Change the Economics of Legal Work

The adoption of legal AI could eventually affect how law firms allocate time across different types of work. Tasks that involve extensive document processing may become faster, allowing lawyers to spend more time on analysis, client communication, negotiation and other activities that require human judgment.

However, the effect will depend on how firms deploy the technology. AI-generated work still requires review, and the cost of maintaining secure systems, integrating existing software and training professionals can influence the overall benefits.

There is also a difference between using AI for relatively routine document tasks and applying it to complex legal questions. The level of human oversight required can vary significantly depending on the consequences of an error and the nature of the matter.

What OpenAI’s Legal Expansion Signals

The launch of Astra for Law signals a broader change in how AI companies are approaching professional markets. The emphasis is increasingly moving from general conversational assistance toward specialised systems connected to industry-specific information and workflows.

For law firms, that means the next phase of AI adoption may involve more than purchasing access to a chatbot. Firms may increasingly evaluate how AI connects with document-management systems, legal research platforms, case information, internal precedents and established professional processes.

OpenAI’s initial rollout to a limited number of law firms also suggests that professional legal AI remains an area where implementation, security and workflow considerations are important alongside model capability.

As competition between AI providers expands, legal professionals will have to assess not only how capable a model is at generating text, but also how securely it handles confidential information, how well it integrates with existing systems and how effectively lawyers can maintain oversight.

The Next Phase of AI in Legal Services

OpenAI’s entry into specialised legal AI places the company alongside a growing group of technology providers seeking to reshape document-heavy professional work.

Astra for Law combines GPT-6 Astra with legal-focused tools, firm knowledge, specialist integrations and professional workflow controls. Its intended role is to help lawyers research matters, analyse documents, prepare drafts and review transactions while leaving legal judgment and professional responsibility with humans.

The broader competition with companies such as Anthropic is likely to push AI providers toward increasingly specialised professional products. For law firms, the important question will not simply be whether AI can complete individual tasks, but whether these systems can be integrated safely and effectively into the complex workflows through which legal services are delivered.

OpenAI’s legal launch represents another step in that transition, moving AI from general-purpose assistance toward deeper integration with the specialised tools, information and professional practices of the legal industry.