ChatGPT: Is AI Reshaping the Legal Industry?

All News

By Attiya Khan

Generative Artificial Intelligence (AI) has been revolutionizing various industries and the legal field is no exception. It is already being used to automate routine tasks such as reviewing contracts, searching through discovery, and conducting legal research. More controversially, though, it is starting to move in the direction of drafting contracts and predicting legal outcomes. While most attorneys acknowledge AI as a friend rather than a foe given the benefits of increased productivity and avoidance of costly mistakes, some players in the industry are resistant to AI integration – and it’s not just because it cuts against their billables.

ChatGPT, a machine learning chatbot created by OpenAI, has been all the chatter in the last few months. It threatens both Google and Microsoft with its ability to generate human-like text responses to prompts inputted by the user. Not only can it provide answers to simple queries, but it can also translate languages, summarize lengthy documents, and analyze sentiment. For example, I can ask it to write me a strongly-worded letter to my neighbor for stealing my packages, and it will do so in less than ten seconds. More importantly, I can ask it to write me a professional memo to a potential client, a discovery demand in a criminal case, or an employment agreement for a hospital physician. This is why ChatGPT is the shiny new thing for every attorney short on time or with a long list of higher priority tasks. It certainly doesn’t provide a final product, but it does provide a workable rough draft, hence why attorneys shouldn’t fear being replaced. Even as machine learning becomes more advanced, attorneys should use the technology to enhance their work because the goal is to produce something better than a machine and human could do alone.

            The critical question is whether law firms and courtrooms are ready and willing to accept the great gift that is generative AI. The short answer is not anytime soon. Despite its sophistication, chatbots like ChatGPT are ridden with legal and ethical concerns. First and foremost, attorneys have a duty to provide competent representation and a duty to communicate. This requires a technical understanding of how chatbots work, what the benefits and risks, as well as intelligible explanation of such information to clients. This is a tall order, especially for senior attorneys who are less proficient in technology. One of the major risks is data bias – the admonished garbage in, garbage out. Unrepresentative or incomplete training data can lead to unfair and prejudicial outcomes, or outright false information. Not to mention, a significant part of attorneys’ work is often trying to find a loophole in the law or a distinguishing element that sets a case apart from its precedent. Existing data will not capture such findings because it may be an issue of first impression, or the facts of a case may be sufficiently unique. Another risk is the lack of confidentiality if client data is stored with third parties (attorneys also have a duty to maintain confidentiality). As we all know, breaking attorney-client privilege can kill a case. Attorneys cannot point the finger at a chatbot when something goes wrong. The looming threat of malpractice is perhaps the strongest deterrent for generative AI in law firms and courtrooms.

            There is one more thing that may weigh against the integration of AI, especially in law firms: the loss the billable hours. ChatGPT is incredibly user-friendly. It is simple and intuitive making it an ideal tool for legal research and document review, especially for non-attorneys. To top it all off, the service is currently free! People with simple legal questions will no longer seek an attorney; they will ask a chatbot whether they have a claim or legal exposure. Even in-house counsel may limit the amount of work they assign law firms because they themselves can review and edit ChatGPT work product. On the bright side, there is more of an opportunity to engage with existing clients and solicit new ones with the time that would’ve been spent drafting up contracts. Attorneys can prioritize client meetings, deal negotiation, litigation preparation, and other in-person interactions. This will fare well with clients who appreciate one-on-one time with their attorneys. Law firms can even use ChatGPT to create marketing materials to attract potential clients. The possibilities are truly endless, as long as users keep the risks in mind and work towards minimizing them.

            Earlier this year, a chatbot built on OpenAI’s GPT-3 was set to represent two defendants in traffic court. The defendants were going to launch a phone application chatbot made by the company DoNotPay that records the proceedings and dictates legal arguments in real-time via smart glasses. This could become a regular occurrence in the courtroom. Consider a world where the cost of an attorney is equal to the price of smart glasses rather than thousands of dollars. This would devastate the already over-saturated legal market. In fact, DoNotPay’s front page literally says “sue anyone at the press of a button”. “Pro se” – in quotes because the chatbot is technically the attorney – litigants will be able to argue with trained and experienced attorneys who have spent hours reading and crafting legal arguments. While these traditional attorneys are fleshing out these arguments in court, the pro se litigant can merely counter any argument presented by the attorney by relying on what comes from their chatbot attorney in mere seconds. The pro se litigant should move cautiously, though, because the chatbot can give an incorrect argument or phrase it in a way that the judge may find offensive. This could lead to reprimands such as fines or being held in contempt, let alone an unfavorable verdict. DoNotPay, however, said it would cover any fines or court fees for clients that use its chatbot services. Unfortunately, multiple state bar associations have threatened prosecution and prison time for the unauthorized practice of law, so the DoNotPay chatbot will not get its day in court just yet.

            Ultimately, generative AI is here to stay, and it is becoming more and more adaptive to the legal industry. It would be silly to turn a blind eye, but it would foolhardy to make an immediate transition, especially in the courtroom. From a forward-looking perspective, though, ChatGPT’s potential is beyond comprehension, regardless of the industry.