Unlocking Legal Efficiency: The Rise of…

Frequent headlines highlight lawyers facing repercussions for including, as one judge noted, “false AI-generated research” in their submissions. The specifics may differ, but the common theme remains: an attorney relies on a large language model (LLM), such as ChatGPT, to assist with legal research or drafting, resulting in fictitious cases being fabricated by the model. The lawyer is often unaware of these inaccuracies until pointed out by a judge or opposing counsel. For instance, attorneys involved in a 2023 aviation lawsuit were fined for filing documents that included these AI-generated fabrications. So, why do they continue this practice?

The primary reason is time constraints and the growing integration of AI across professions. Legal research platforms, including LexisNexis and Westlaw, are now equipped with AI features. For busy lawyers, AI may appear to be a highly efficient tool. While many attorneys do not use ChatGPT directly for drafting legal documents, they increasingly depend on it and similar models for research purposes. However, many of these professionals lack a clear understanding of what LLMs are and how they function. One sanctioned attorney in 2023 believed ChatGPT to be a “super search engine,” only to discover, following the submission of flawed citations, that it is closer to a generator of seemingly random phrases, sometimes providing accurate information, while at other times producing convincing yet incorrect content.

Andrew Perlman, the dean of Suffolk University Law School, points out that the majority of lawyers are using AI without facing issues, with those caught providing false citations being exceptions. “While hallucinations are a legitimate concern that attorneys must take seriously, that doesn’t diminish the significant benefits and applications these tools can offer in delivering legal services,” Perlman noted. Legal databases and research tools like Westlaw are integrating more AI features as well.

A survey conducted by Thomson Reuters in 2024 found that 63% of lawyers reported having used AI previously, with 12% stating they use it regularly. Many respondents indicated they utilize AI for drafting case law summaries and researching statutes, orders, or sample language. The surveyed lawyers view AI as a time-saving resource, with half of them highlighting the exploration of AI’s implementation in their work as a top priority. “A competent lawyer acts as a ‘trusted advisor,’ not merely as a document creator,” commented one respondent.

Nevertheless, numerous recent cases demonstrate that not all documents produced by AI are accurate, with some being entirely fabricated.

In a notable recent case, attorneys representing journalist Tim Burke, arrested for publishing unreleased Fox News footage in 2024, filed a motion to dismiss on First Amendment grounds