by Steve LewickyUsing Artificial Intelligence in litigation and legal disputes
Artificial intelligence is becoming a big part of everyone’s business and professional life, including when legal disputes arise. When my clients become involved in business or real estate disputes, or other litigation or arbitration, they often now consult AI agents before talking to an attorney. This desire by clients to empower themselves with knowledge and avoid unnecessary legal expense is understandable. AI can be very useful when you have questions about the law or are in a dispute – but while AI is useful and time-saving for some tasks, it can be very unreliable for others. Knowledgeable clients get the best results when they treat AI as a powerful organizational and analytical tool, but not as a substitute for an experienced attorney.
Business disputes can generate enormous amounts of information, such as contracts, emails, text messages, invoices, spreadsheets, accounting records, photographs, correspondence, deposition transcripts, and other documents. One of AI’s greatest strengths is its ability to process and organize large amounts of information quickly. A business might have several years of invoices and payment records, for example, and AI can sometimes help identify payment patterns, total up amounts paid, compare two sets of financial records, or identify transactions that appear to be missing from a set of records. AI may also be useful in reviewing a collection of emails and identifying communications concerning a particular subject, person, transaction, or period of time. Sometimes my clients use AI to summarize a lengthy contract, organize a chronology of events, extract dates from correspondence, identify the principal subjects discussed in a group of documents, or create a table comparing different versions of an agreement. These tasks can be valuable because they allow a lawyer and his or her client to spend less time performing labor-intensive review work, and more time analyzing what the documentary information means. In many cases, thoughtful use of AI can reduce overall legal expense.
AI also can be helpful to a client in preparing for meetings with counsel. A client that organizes the history of a dispute into a concise timeline, identifies important documents, and lists the major questions that need to be addressed, will usually make better use of the time spent with an attorney.
AI is considerably more problematic, and potentially harmful, when a client goes from using it to organize information, to providing legal opinions, strategy or conclusions. AI agents might, in some circumstances, provide a useful starting point for identifying legal issues or suggesting subjects that should be researched – but they frequently make mistakes when describing statutes, court rules, and judicial decisions. There are many examples in published court decisions and other public sources of AI providing incorrect quotations, misunderstanding court decisions, relying upon outdated law, and sometimes giving citation to court opinions that do not even exist. A person using AI for any form of legal-related work should never assume that a legal proposition is correct simply because an AI system states it confidently, or provides what appears to be a legal citation. Every significant legal authority as to be independently checked against a reliable legal research source.
This problem is particularly important in litigation. The outcome of a lawsuit or arbitration may depend upon relatively subtle distinctions – for example, which court issued a particular decision, whether that decision is binding or merely persuasive, whether the case has subsequently been limited or overruled, or whether a statute or court rule has recently been amended. AI does not consistently make those distinctions correctly.
AI can generate language very quickly, but that does not mean it is particularly good at drafting pleadings, motions, briefs, or other papers to be filed with a court. Effective legal writing requires more than professional-sounding prose. A good lawyer decides which facts should be alleged, which facts should be emphasized or omitted, which causes of action or defenses are legally available, what must be preserved for later proceedings or appeal, what evidence can actually be proven, and what arguments are strategically wise. A seemingly minor sentence in a complaint or motion may have consequences later in the case, and an allegation may constitute an admission. Unnecessary legal theories may complicate the litigation discovery process. An argument made too broadly may undermine another position. In addition, pleadings and papers filed with a court must comply with the procedural requirements of the particular court in which the case is pending. At the most basic level, AI is not really capable of creating, evaluating or pursuing a well-considered litigation strategy. It can produce a document that looks like a good legal pleading, but it cannot tell you with any reliability what pleading should be filed. AI can be a tool for brainstorming, checking consistency, working through large quantities of information, etc., but the judgment exercised in preparing the final court filing remains fundamentally the lawyer’s responsibility.
Clients get the greatest benefit from AI by using it to help organize and understand their own information. For example, if you have a complicated payment dispute, AI might be useful in organizing invoices and payments into a spreadsheet. If you have a lengthy chain of correspondence, it may help identify the primary communications. The resulting work should then be reviewed with counsel. Your lawyer can determine which facts actually matter legally, whether additional documents are needed, what information may be privileged or confidential, and how the information fits into the overall strategy of the case.
Finally, it is very important to exercise caution before uploading confidential, privileged, proprietary, or personally identifiable information into any AI service. The many AI platforms have different policies concerning storage, retention, and use of information. Before providing sensitive materials to any AI system, a you must understand how that information will be handled by the AI service. Once you upload documents or data, the information will be subject to use and sharing consistent with whatever terms and policies apply to the particular AI service – and may also be dependent upon what level of service you have selected. As a rule of thumb, you should not upload to a “free” AI tool any data or information that is in any way confidential or contains personal information. Whether a “paid” service or participation level will adequately protect your information depends on the nature of that service’s user agreement and policies. It would be safest to discuss these matters with counsel before uploading any confidential or personal information to an AI system.
Properly used and with sufficient privacy and confidentiality safeguards, AI can be an excellent assistant. It is fast at organizing information, can be useful for analyzing data, and sometimes is remarkably effective at summarizing complicated records. AI remains capable of significant factual and legal errors, however, and its apparent confidence should never be mistaken for reliability. The best approach is to assign AI the work it does well and reserve legal judgment, litigation strategy, legal research verification, and final court filings for the attorney responsible for the case.
