ChatGPT and Law Firm Management: Will AI Replace Lawyers?
Even a year after GPT-4 took and passed the uniform bar exam – placing in the 90th percentile – the reverberations remain. A portion of the legal profession seemed worried, while another seemed intrigued. Managers of law firms have largely taken a pragmatic approach. And yet, the question still lingers.
Will AI replace lawyers? Or, for the fatalistic, when will AI replace lawyers?
You can find hundreds of articles, opinion pieces, and studies. We’ve talked about AI in law firms already. Now, we’re going to hit the question head-on.
Use of AI in the Legal Industry
From law students writing reports and even answering exam questions with AI-generated responses to practicing attorneys being reprimanded for submitting a brief authored by AI, the technology is already seeping into the practice of law. This is good news and bad news in a few ways.
The Benefits of AI in Legal Practice
The foundation of nearly every brief, argument, and even novel legal theory is knowledge, specifically statutory and case law. And it’s not just the law. It’s the order of the decisions and legislation that matters. AI is an immensely powerful and capable information aggregator. It can also recognize and recreate patterns.
In this way, an AI can churn through a hundred years of available data and produce an answer to a prompt in mere seconds instead of hours or days of legal research. Legal AI tools can also produce fundamental legal documents if properly prompted and given the criteria for them. The greatest benefit, then is that of time saving.
The Drawbacks of AI in Legal Practice
The short version is this: information access and accuracy. Asking an AI without actual access to case law or the data on Westlaw or LexisNexis inevitably invites fabrication, and there are many well-documented cases of AI-generated text containing factual errors. ChatGPT and its ilk are based on predictive text. They’re not analyzing and understanding the spirit of the law or its intent.
There may also be questions of plagiarism and intellectual property rights at times. U.S. copyright law holds a human element essential to protecting works, so anything produced by an AI can’t presently be copyrighted, which begins to matter when you think of AI-generated legal documents, templates, and even computer code.
AI vs. Lawyers: Competitors or Collaborators?
Current prognosticators see the legal world and its participants in flux. The demand for legal professionals has grown above the general workforce, but the possibility of AI replacement looms especially large as well. Legal assistants and even some paralegals may be right to be wary.
AI tools are not likely to replace lawyers, and we’ll get to the reasons later. Not competitors, then, at least not yet. However, artificial intelligence can make their jobs easier. Remember that AI tools are only as smart as the information they have access to, and even that intelligence is limited in ways that aren’t likely to be overcome. Still, the processing power and ability to thoroughly sift through data to answer prompts has many uses.
Smart lawyers are already putting AI to work in novel ways. Don’t fear replacement – fear the edge competitor firms may get from using AI to get the upper hand.
Why AI Will Never Replace Lawyers
Human lawyers will, in our estimation, always be necessary to provide quality legal services. In their current forms, at least, AIs suffer from a few fundamental limitations that make them unsuitable for the practice of law – whether they can answer bar exam test questions accurately or not.
And if the well-rounded and currently newsworthy analysis of the topic doesn’t sway you, consider the “opinion” of ChatGPT itself. We prompted it with the following question: Do you think you can ever replace a human attorney, and if not, why not? Here was its answer:
While AI technologies have advanced significantly in recent years and can perform various tasks previously handled by humans, it is unlikely that AI will completely replace human attorneys in the foreseeable future. Here’s why:
Complex human interactions: Law involves not only knowledge of legal principles but also complex human interactions, empathy, and emotional intelligence. Human attorneys can understand their clients’ needs, emotions, and unique circumstances, which is difficult for AI to replicate fully.
Legal ethics and judgment: Legal cases often involve ethical considerations and require judgment calls. Human attorneys must navigate gray areas and apply the law to unique situations, something that AI, as a machine learning model, may struggle with.
Adaptability and creativity: The practice of law often requires creative thinking and adapting strategies to different cases and scenarios. Human attorneys can think outside the box and come up with innovative solutions, which is challenging for AI, as it relies on patterns from past data.
Courtroom presence: In litigation and trials, human attorneys must argue their cases in front of judges and juries, and their persuasive abilities, charisma, and presence can have a significant impact. AI lacks the physical presence and emotional expressions that can influence human decision-makers.
Legal responsibility and accountability: The legal profession comes with a high degree of responsibility and accountability for the outcomes of cases. While AI can assist with research and analysis, final decisions and responsibilities often lie with human attorneys.
Confidentiality and trust: Clients often share sensitive and confidential information with their attorneys. Trust and confidentiality are essential in legal matters, and many clients may prefer dealing with human professionals they can personally interact with.
Breaking Down the AI Answer
First off, every AI-generated answer comes with a disclaimer from artificial intelligence that it cannot provide legal advice. Using one without doing one’s own research, due diligence, and fact-checking is just lazy legal work – and legal departments and law schools are already on the lookout. So is the legal system. Here are a few caveats to AI and additional insight into its answers.
1) Predictive text generation is not thinking. It’s pattern matching. The AI doesn’t have an “idea.” It is scanning through available data and recognizing patterns that it will then amalgamate into an answer. If you were to thoroughly search for the ideas in each of the numbered answers above, you might find similar text strewn in multiple articles. Nothing new is created – just a novel way of presenting what already is. And if “what is” is effectively a mass of opinion, then a predictive AI is simply parroting that opinion.
2) Humanity is an essential part of the law, and AI can’t play it. An AI is currently incapable of suffering a personal injury or a divorce. We don’t hear judges fearing for their positions on the bench. And an AI isn’t ever likely to sit on a jury either. So, if it isn’t going to be a client, won’t replace the judge, and can’t be on the jury, why would anyone think an AI could replace the other human in the simplified process: the lawyer?