The Impact of ChatGPT on the Legal Industry: Balancing Workflow Benefits, Security Concerns, and Job Loss Fears

The buzz around ChatGPT securing 100 million users just two months into its launch was second to none. The OpenAI-developed AI chatbot was a revolution in the truest sense of the word. You could type in any question and get an almost instantaneous (and quite reasonable) response. From mundane queries to complex legal paperwork, ChatGPT could do it all.

Of course, as the hype settled down and the technology's pervasiveness became more noticeable, industries soon began to take notice of the workflow disruptions it could accrue. Around the tail end of March, Goldman Sachs reported that about 300 million jobs worldwide could be lost due to AI-driven automation. This sobering reminder to the world soon created a sense of urgency within corporations to take advantage of this technology and its potential. On the flip side, however, it also ushered in a sense of fear and trepidation.

As it stands today, the notion of AI's potential and the associated job loss concerns transcend beyond the corporate and venture capital world. It has entered into the mainstream lexicon across all industries, many of which will be directly impacted by AI. To that end, this blog post will serve as an enlightening exploration of the impact of ChatGPT on the legal sector and what it may mean for legal professionals going forward.

Andrew Perlman, in his article for the Harvard Law School, writes that "ChatGPT may portend an even more momentous shift than the advent of the internet." And that bodes well when you consider the potential implications of the technology for the legal profession. Consider this; a lawyer's day-to-day work involves a great deal of repetition, such as writing briefs, researching precedents, and preparing documents. These tasks can be time-consuming and tedious. No doubt, they're crucial elements of the job, but ones that leave little room for creativity or an opportunity to think strategically and outside the box.

ChatGPT provides a solution here, allowing lawyers to automate the tedious aspects of their work while freeing up their time to devote to more meaningful tasks. That means they can focus on the creative, strategic aspects of the job – those that truly require a human touch. Consider this example from Perlman's review of the chatbot. He asked the bot to write a research paper about ChatGPT's potential applications in law. The bot responded by saying that it could help with legal research, generating briefing documents, relaying generic legal information to users, and even conducting thorough legal analysis based on the set policies.

A recent Thomson Reuters article explored the role of ChatGPT (and generative AI in general) in the legal framework by analysing the use cases, concerns, and limitations addressed in the recent Legalweek event. In the event, Danielle Benecke, founder of a law firm, outlined that "firms and other enterprises have been sitting on this unstructured data forever" and generative AI technologies could be viable in "[supercharging] the firm's most valuable pre-existing service lines."

So, optimism is there, and the potential of generative AI technologies in legal frameworks is apparent. Yet, there are still many concerns that must be addressed before ChatGPT. There's the discussion around:

• AI hallucinations that could lead to wrong decisions
• AI-driven decisions that could lead to unfairness
• Ineffective linguistic adeptness when kept relative to the sophistication of the legal language
• Lack of precision in decisions due to constraints on the interpretation of legal concepts

Favourably, it looks like these limitations are also being taken into account. A judge in Cartagena, Colombia, integrated ChatGPT into his workflow for legal decisions across hearings. He said that while ChatGPT wasn't really a replacement for human decision-making, it served well to "optimise the time spent in drafting sentences, after corroborating the information provided by AI."

What does this tell us? AI-based tools have the potential to provide a lot of support in understanding legal documents and generating them in the first place. And that begs the question - how is this being received on the ground?

We recently conducted a comprehensive survey of 93 law firms. We asked them about their experiences with AI tools (specifically ChatGPT), how they have been employing them, what are their perceptions and expectations, and the current state of AI-human workflow across these legal firms. The responses were positively in favour of the usage of AI-based tools such as ChatGPT.

• Almost 30% of them reported that they are using ChatGPT as one of their work tools.
• In fact, four legal firms revealed that their paralegals (legal assistants) have been spending 70% of their time working with ChatGPT.
• On average, legal firms were spending 2-3 hours a day using ChatGPT.

If we map these findings to Gartner's 2022 predictions about legal technology's evolution, there is a strong indication that AI-based tools will get more and more acceptance across the legal industry. For example, Gartner outlined that:

• 50% of the legal work by 2024 will be automated
• 25% of the corporate legal application spending would be associated with non-specialist technology providers
• By 2025, legal tech budgets would increase by 3X when compared to 2020

In March 2023, researchers from Princeton University, UPenn, and New York University revealed the most exposed industrial domains to the rising influence of AI. The study put "legal services" at high risk to the exposure of advanced language models. Researchers included profiles like lawyers, paralegals and legal assistants, and legal secretaries to substantiate this argument.

Cuing back to the above Goldman Sachs report, the investment bank also affirmed the high-risk exposure — outlining how 44% of the legal workloads could be automated. So, as ChatGPT continues to gain acceptance in the legal industry, it will have a significant influence on how the legal job market plays out.

On the one hand, legal firms may leverage it to reduce costs for internal operations. However, on the flip side, they could proceed with hiring fewer employees. Of course, that doesn't mean that lawyers will become obsolete. But, it does reflect upon the market shift. Let's explore these concerns more profoundly:

• Job Loss - The Low-Hanging Fruit

  Paralegals and legal assistants who are responsible for collecting, analysing, and sifting through large volumes of legal documents and data could likely be the most to be affected. Today's large language models (LLMs) and the technologies developed on top of them (like GPT 3.5 and GPT 4 that are powering ChatGPT) are the most "cutting edge" in terms of AI-based tools.

  Given that these models can be readily implemented on top of the current paralegal systems, it is likely that the paralegals in law firms could find it challenging to compete with ChatGPT, especially for repetitive tasks.

  This trend is, in fact, slowly transpiring across geographies. Consider the example of the US. The US Bureau of Labor Statistics outlines that opportunities for "office and administrative support occupations" have been declining over the years. By 2031, they expect about 880,800 fewer jobs to be advertised in the market.

• The Concerns of Bias and Inaccuracy

  Recently when OpenAI released their most advanced language model yet (GPT-4), they outlined how it was 40% more accurate than the previous iterations. This was in line with the objections that were raised around AI hallucinations and bias in chatbots.

  The accuracy of AI-based technologies, especially language-based ones, has improved at a mind-boggling rate over the past few years. However, an utterly sophisticated and regulated industry like the legal sector brings new concerns - the legal language and concepts can often be very complex, making it hard to imbue ChatGPT with the same level of complexity.

  It is evident that there is a lot of preparation work that goes into legal research, so the lack of linguistic adeptness could lead to inappropriate and biased decisions. And it's not only about ChatGPT's linguistic abilities. The possibility of AI-based technologies chauvinistically being used on people is also worth considering.

  So, of course, there's this ray of hope as well that transpires from the complex nature of the industry and the relatively bland and straightforward outputs handed by generative AI technologies like ChatGPT.

The prowess and inefficiencies of generative AI technologies have been on full display this year. Favourably, the legal industry has been keeping abreast of the developments, something that a Thomson Reuters survey of legal professionals across the US, UK, and Canada substantiates. In it, about 15% of the legal firms issued warnings about the use of ChatGPT at work. In fact, 6% had banned the usage altogether.

However, the same survey also suggested that more than 80% of the firms recognise AI's potential to be integrated into some workflow in the legal ecosystem. So, amid this uncertainty about the technology's viability, how should legal firms plan for the future and keep on top of the developments?

1. Keep an Eye on Black Boxes

   As the use of ChatGPT and other generative AI-based technologies continues to expand across the legal industry, it's likelier that some inappropriate and biased assertions will be made. Legal firms must be on the lookout for biases and inaccuracies in the outputs that are being delivered by ChatGPT. It is essential for legal personnel to scrutinise these outputs and make corrections where necessary. After all, transparency is a key component of good lawyering.

2. Prepare for the Unknowns

   Generally, regulatory and compliance frameworks evolve at a slow pace. While these frameworks sometimes result in some adverse outcomes, they also introduce sustainability to business operations. Law firms must be prepared for the unknowns that lie ahead in this regard. They can do this by ensuring that they are in sync with the application of ChatGPT and other AI-based technology in their firm.

3. Develop Holistic Visibility into Day-to-Day AI Usage within the Firm

   While surveying the legal firms, we also found that a host of these organisations have been able to understand the patterns of AI usage across their departments by analysing data collated by CleverControl .

Now, as we move closer to the time when ChatGPT and generative AI-based technologies become ubiquitous in the legal industry, firms can leverage their insights from CleverControl to maintain holistic visibility into the usage of ChatGPT and other generative AI-based technologies. They can even strategise for:

• How to humanise the process that entails utter sophistication
• What are the "sweet spots" and "black spots" in the ChatGPT usage across departments
• What are the most applicable use cases of ChatGPT across departments

Altogether, these findings suggest that AI-based technologies hold the promise of automation across the legal industry. It is imperative that firms stay ahead to extract maximum value from these technologies and empower their employees in the process.