AT&T and OpenAI Are Building a System That Takes Legal Requests in Plain Language
Employees would submit legal requests in plain, everyday language, and AI would classify them and pull up relevant past matters. The legal decisions still belong to attorneys.
OpenAI said on October 21 (local time) that it is working with AT&T to build a system that supports the telecom company’s in-house legal work. The core idea is to take in legal requests submitted by employees and surface the knowledge relevant to them. The effort builds on work already under way at LegalEdge, AT&T’s AI team within the legal organization.
AT&T’s in-house legal department includes several hundred attorneys and legal professionals around the world. The collaboration is focused on making it easier for them to find legal knowledge. The goal is to support everything from the moment a request arrives and its contents have to be understood, through the step of looking back at similar matters handled in the past.
How legal requests get handled
Requests come in using ordinary, everyday language. Whoever needs legal support can simply describe it the way they would speak or write it. The system then classifies and prioritizes what it receives, indicating what kind of request it is and where it falls in the queue.
Attaching related material to each request is also part of the plan. The system would link each request to relevant precedents, research, and past matters the legal team has handled. Rather than merely displaying the contents of a new request, the aim is to let the legal team see it alongside matters they have already dealt with and the knowledge they have accumulated.
For the legal professional assigned to a request, the system also presents background on earlier matters and a recommended first action. Beyond retrieving relevant documents, this would help the assigned professional understand the context of earlier matters and begin reviewing the request right away. The concept ties intake, classification, document linking, and a suggested first step into a single workflow.
Who makes the legal calls
Legal decision-making stays with attorneys, OpenAI said. The first action the system proposes serves only to assist the assigned professional’s review. Even when AI organizes the request and the supporting material, the legal judgment on any given matter belongs to a lawyer.
The system is for AT&T’s in-house legal team, and it is still at the planning stage. What changes is the way a request is first received and the way information needed for review is gathered. The emphasis is on letting the person handling a matter examine the request and related precedents in one place.
The direction AT&T and OpenAI have laid out is to use AI to help professionals quickly obtain the background they need to make a judgment. What remains to be seen is how the concept works out in actual intake operations, and how much it changes the process of finding material and preparing a review.