Artificial intelligence (AI) has been part of the legal field for a while. It has grown from litigators using it to rapidly review mountains of discovery documents to many additional use cases today. AI-enabled legal technology is pervasive, though not all law firms are using it. Even the latest headline-generating AI technology—ChatGPT—has made its debut in legal.
Let’s review exactly what AI is and how lawyers have been using the technology. We’ll also look at the emergence of ChatGPT and how it might impact the legal profession.
Artificial Intelligence & Legal Use Cases
Artificial intelligence seeks to mimic human intelligence using a computer. A machine is fed instructions (algorithms) on what to find, and off it goes to search and serve-up potential “hits”. For example, many people encounter AI in their smart devices when they say, “Alexa, play my Chill Playlist.” In the legal context, this might sound like, “Find me all mentions of MRI in these medical records.”
Legal AI is a tool lawyers use to review and analyze masses of data. AI also helps a legal practice automate mundane and administrative tasks. Here are a few ways AI is currently being used in the legal field.
TAR, also known as predictive coding and computer-assisted review, was one of the early AI applications. With TAR, litigators reviewed a sampling of documents for responsiveness and hot issues, then fed the machine instructions on how to identify these in the full data set. Later known as machine learning (ML), this kind of AI learned as it went along and got smarter about what the user wanted.
It’s important to note that lawyers always review what the machine finds—confirming it’s what they need and not privileged. Although some AI can identify privileged material too. Using TAR as a workhorse drastically reduced discovery review time and cost.
Judge Andrew Peck was the first to rule that courts will permit the TAR for document review in 2012.
Today, AI also drives many of the legal research tools lawyers use. AI doesn’t require the painstaking Boolean logic we lawyers used to peck out. With AI, lawyers and paralegals can quickly search large databases of statutes, regulations, jurisdictions, and case law.
For some litigators, using AI to predict case outcomes based on judges’ and lawyers’ past behavior and rulings has become routine.
One area where legal AI is catching on is practice management. Artificial intelligence powers many of the legal tech products that automate invoice generation, payment reminders, conflict checks, and client onboarding tasks. You can use this tech to sign up new clients faster and get paid on time.
Calendaring and scheduling appointments are also assisted by AI. For example, one product connects inboxes and calendars. To set up a meeting you’d send an email request, cc’ing the AI “bot”. Within an hour, you’ll get an email listing the times and dates you could meet.
Managing All the Documents
Lawyers generate tons of documents, including briefs, motions, agreements, client communications, pleadings, engagement letters, etc. Documents are what we do. Document management systems with AI help lawyers rapidly find documents or emails they need using the client name, document types, dates, and other filters. This reduces the frustrating time you spend searching for documents.
AI tech can also predict what client/matter folder an email should be stored in within the document management system. This way you have all the documents and emails for a matter automatically stored in one, easy-to-find place.
A lawyer can create a pleading or stock purchase agreement by simply filling in fields in a form. The AI system will automatically generate a document from a library of smart templates. You’ll want to review the document to make any needed changes, but this AI document generation gives you a jump start.
One important note: You will have to invest some time in creating the template library.
Contract Lifecycle Management (CLM)
Many CLM products use AI to help lawyers manage all aspects of contracts. These systems can create initial contract drafts with the relevant standard clauses you’ve developed for a client. The AI can also scour contracts sent to you and compare the proposed clauses with your standard clauses to identify what you’ll want to negotiate. Key milestones and renewal/expiration date alerts help you manage the full lifecycle and avoid having terms you wanted to change auto-renew without your awareness.
AI in the Criminal System
State criminal judges making pre-trial detention, sentencing, or early release decisions have used AI technology to assess the risk that a defendant or convicted person will commit a crime again (recidivism). However, critics say that conviction and arrest data used to train the machine learning AI are biased and unfair to people of color.
Transformative GPT AI Technology
There’s a transformative AI technology screaming from the headlines in 2023. ChatGPT “chats” have a conversation with a user via text and then create content in response. GPT stands for “generative pre-trained transformer.” OpenAI created this generative AI and trains it to use internet data to generate virtually any type of text.
This super AI can write music, jokes, puzzles, emails, and articles. It can also extract information from contracts, analyze text for expressed sentiments, find bugs in code, create websites, and much more. One is left with the question—is there anything generative AI can’t do?
A word of caution—ChatGPT is an infant. It remains in the research and feedback-collection phase as of February 2023. The product itself states that ChatGPT “may occasionally generate incorrect information” and “may occasionally produce harmful instructions or biased content.”
Who is Using Legal Generative AI?
Generative AI is in the very early stages of incorporation into the practice of law.
OpenAI backed a startup that is using GPT to develop a tool for legal AI generation—HarveyAI. A very large London law firm says 3,500 of its lawyers are using the tool, testing it for simple legal questions and document and email drafting.
In a Columbian court, a judge recently used ChatGPT to both search for precedents and draft his ruling. He maintains that a judge must always corroborate the information he uses—whether a legal secretary or AI supplies that information. He also believes ChatGPT will help reduce Columbia’s large case backlog.
Several legal tech startups and existing legal tech companies are also experimenting with GPT-3 features or other large language models. These companies provide AI legal assistants, legal research, knowledge management, and drafting technologies.
Where is Legal AI headed?
In general, most lawyers are not rushing to adopt AI. They worry that an AI mistake could cost them a case, or worse, their career. The fear that AI will replace lawyers also still festers.
Let’s not forget though that even lawyers who were extremely skeptical of TAR in 2012 use it today for large document sets. The efficiency and productivity gains of AI are simply too large to ignore. Lawyers at large law firms and corporations have been putting their toes in the AI water with document management, invoice generation, and contract management for the last few years.
For ChatGPT in legal, concerns about its accuracy and copyright issues will likely limit adoption in the near term. Firm leadership will stew over the lawyer and staff time invested in fact-checking ChatGPT outputs, so are the risks worth it? In the not-too-distant future, however, law firms and corporate legal departments will use AI advances like ChatGPT to perform work once delegated to junior lawyers and legal assistants. Ultimately, AI tools will free lawyers and staff up for more strategic work and continue to change how lawyers practice law.