The Don'ts of Using AI in Legal Work

The integration of artificial intelligence (AI) into legal practices has opened up a myriad of possibilities for efficiency and innovation. You can do many things with AI: perform basic legal research; draft briefs; correct or improve sentences and paragraphs; outline arguments; conduct document review, analysis, and summarization; review contracts; analyze intellectual property; and much more! However, lawyers must observe critical boundaries and ethical considerations when leveraging AI.

Here are the top five things lawyers should not be doing with AI

Do Not Violate Client Confidentiality When Using AI

Lawyers must make sure their use of AI does not compromise client confidentiality. Sharing sensitive or confidential information with AI systems, especially those that are not secured or those operated by third parties, can lead to breaches of confidentiality. Lawyers must make sure any AI tool they use follows our stringent confidentiality standards.

There are two actions you can take to help maintain your client's privacy:

1. Don't use clients' confidential information when using AI. Client confidential information includes personal identification information, financial records, strategic business plans, and private communications regarding legal advice and case strategies. If you don't put it into the system, the AI can't use it or leak it.
2. Check the Terms of Service (TOS) of the AI you are using. The TOS should detail how the AI provider collects, uses, stores, and potentially shares data input into the system. This could easily include sensitive client information. Attorneys must understand these terms to help us assess the risks of data breaches or unauthorized access. For example, the terms might reveal whether the AI provider claims any rights over the data processed through their system, which could compromise the confidentiality and ownership of client information. Additionally, the TOS may include information about the jurisdiction and legal frameworks governing the data, which is important for compliance with applicable privacy laws and ethical obligations. Last, by understanding the terms, attorneys can identify the security measures the AI provider implements to protect data, ensuring they align with our profession's stringent confidentiality standards.

Common sense tip: If you are using a "free" version of an AI product, there is a good chance the information you provide it will not be confidential.

Do Not Have AI Write Your Brief and Then Not Review It

1. AI does not replace a lawyer's judgment. A lawyer should not let AI write a brief without critically reviewing it. Despite AI's advanced capabilities, it lacks the nuanced understanding, ethical judgment, and accountability that a lawyer brings to legal writing. AI-generated texts might miss subtle legal nuances, misinterpret facts or laws, or fail to tailor arguments to a case's specific context and jurisdiction. AI is not a human being (at least not yet) and cannot replace human judgment.
2. AI is known to hallucinate or make stuff up—such as case citations. Lawyers should be concerned with AI "hallucinating" or generating factually incorrect or nonsensical information, a very good reason why you don't want AI to write your brief without review. AI, particularly language models, can sometimes produce content that seems plausible but is entirely fabricated or irrelevant to the context. This includes making up case citations! In legal documents, where precision, accuracy, and reliability of information are paramount, such errors can lead to misrepresentations, weaken a case, or even result in sanctions against the lawyer for submitting misleading or incorrect information to the court.

Just as a lawyer would review a paralegal's work to ensure accuracy, coherence, and strategic alignment with the case objectives, a lawyer must scrutinize an AI-generated brief to ensure it meets legal standards and ethical obligations while effectively advocating for the client's interests. This review process is important for maintaining the quality and integrity of legal work and fulfilling the lawyer's professional responsibility to provide competent representation to their clients. Your credibility as a lawyer may be on the line!

Misrepresenting AI Capabilities and Their Use to Clients

It's unethical for lawyers to overstate the capabilities of AI in order to impress or mislead clients. This includes suggesting that AI can guarantee certain outcomes or replace the nuanced advice that comes from years of legal training and experience. But whether lawyers need to tell their clients they use AI in their work could be up for debate.

Lawyers might need to tell clients if they use AI in their work, maybe even putting it in the contract. This is important for being open and ensuring clients agree and understand what's happening with their information. If AI is a big part of the work, such as researching or reviewing documents, clients will more likely need to know. Informing the client helps build trust and ensures clients are satisfied with how their cases are handled, especially with regard to keeping information safe and the quality of the work.

But one might question the necessity of explicitly disclosing AI use. Lawyers typically do not discuss the specifics of their word processing programs, grammar review tools, spell checkers, internet browsers, or even the online legal research libraries they utilize. Is the use of AI fundamentally different in a way that warrants specific disclosure or should it be considered just another tool in the legal professional's arsenal, akin to these other technologies?

In the end, you may want to err on the side of caution!

Lawyer Must Be Aware of Potential AI Bias

AI systems learn from large amounts of data and algorithms, which can sometimes include biases. These biases can lead to unfair results. This is especially important in areas such as criminal justice, hiring, and family law, where AI might wrongly influence decisions on things like sentencing, job selection, or custody. Lawyers using AI must carefully check these tools for any biases or mistakes. They should look at how the AI works and where its data comes from, as this data might have built-in biases. Using AI without understanding its workings and data sources can lead to biased results in legal advice, legal analysis, and legal decisions. Accepting what AI says without question can harm the fairness and trust in the legal system.

A Lawyer Is Ethically Obligated to Stay Informed about and Maintain Competence in AI and May Even Be Required to Incorporate AI into Their Legal Practice

AI can significantly enhance the practice of law. As the legal profession evolves with technology, lawyers have a duty to maintain competence in their practice. See MRPC 1.1 Comment 8. This includes understanding the benefits and risks associated with AI. Ignoring the advancements in AI, refusing to understand its implications on the legal profession, or even not using AI in the practice of law can result in a failure to provide competent representation to clients.

This could lead to bar disciplinary actions. And nobody wants that!