Law School Externships Without AI Policy Is Risky for Students

Law schools and professors have varying approaches to teaching and allowing the use of artificial intelligence in the classroom, and as a result, students’ experiences and exposure to AI in legal practice ranges widely.

With AI being integrated in almost every industry, workplaces hosting law students for an externship must have a clear AI policy to help students understand site specific expectations about its use and avoid any misuse or potential character and fitness issues.

A Clear Policy

Summer externships provide opportunities for experiential learning, primarily with governmental or nonprofit entities. These entities often have limited resources, and it can be tempting for students to use AI to increase efficiency.

Sites will differ on whether AI use is permissible. By issuing a clear AI policy that’s explained to the students, it ensures students understand expectations when performing legal research, drafting legal documents, and interacting with clients.

Law school externship programs should require externship sites to have an AI policy and ensure the policy is explained to the students. Unlike summer associates at law firms, externships are generally unpaid and instead provide students with law school credit. Many firms have a policy and trainings around AI, and even with that, there are notable examples of attorneys filing material that contains generative AI produced hallucination content.

Externship sites, such as judicial chambers and local government offices, may either have a prohibition against using AI or may not have the resources to purchase AI integrated legal research and writing tools beyond personal ChatGPT or Claude accounts.

The lack of clear expectations around AI usage could lead to disclosure of client confidential information, waiver of privilege, or inclusion of hallucinations in work product, any of which would be detrimental to the law student and their summer externship employer.

Additionally, standards regarding AI usage in legal practice are rapidly developing. The Florida Supreme Court recently amended the Florida Rule of General Practice and Judicial Administration to require a representation when documents are filed in a Florida court that “the legal authorities identified exist and are accurately cited.”

Last fall, the Illinois Attorney Registration and Disciplinary Commission issued the “Illinois Attorney’s Guide to Implementing AI.” As externship sites work through these jurisdictional requirements, they should take steps to ensure students know how to navigate proper AI use.

Best Practices

First, an externship site must decide whether to allow students to use AI in legal work and to what extent or prohibit students from using it. Next, the site must develop a policy that reflects this decision. Finally, this policy must be communicated to the students at the beginning of the externship.

An effective policy first defines artificial intelligence and provides how different forms of AI may or may not be used in that workplace. While lawyers have been using AI for years, such as for document review and e-discovery, the use of generative AI is more challenging because large language models create new content. Site hosts should specify if and how law students can use the following types of tools:

• Public-facing generative AI (e.g., ChatGPT, Claude)
• Enterprise licensed generative AI
• Enterprise licensed legal tech

Additionally, an effective policy will provide detailed guidelines for what acceptable AI use means. Are lawyers allowed to use it for reading and summarizing, researching, brainstorming, writing, drafting, editing, creating tables, and organizing information?

While defining the tools that lawyers are permitted to use can impact issues of privilege waiver or likelihood of hallucinations, the way in which AI is used can also impact these considerations. Using AI to summarize a public filing opinion has its drawbacks, but it’s unlikely to involve sharing confidential or privileged information.

Lastly, the workplace should consider any additional parameters and the reasons behind them. If AI use is permitted, one common requirement is an absolute ban on disclosing confidential information. Client names, specific fact patterns, opposing counsel details, and personal information should never be entered into an unapproved AI system. Understanding the reasons for the policy will make it more likely that externs and lawyers at the workplace will adhere to the policy and apply its principles in new contexts.

Ideally, the practicing attorneys at the workplace use AI in compliance with its policy. It will be hard for students to adhere to an AI policy if the attorneys with whom they are closely working are not adhering to it. Disregarding the policy may include using a personal ChatGPT account for brainstorming arguments or initial legal research.

Students need to be trained on the externship site’s policy especially if there’s an expectation to use AI. If the AI policy prevents the use of generative AI, then there should be training for students on tools that don’t have integrated AI.

Many attorneys and legal professionals struggle with properly using AI. Establishing a policy doesn’t prevent AI misuse, but sets the tone by acknowledging it exists, setting expectations, and starting the conversation for law students as they learn.

Both the law school and the externship site need to establish open lines of communication where students feel comfortable asking questions about AI misuse and subsequent consequences around it.

This article does not necessarily reflect the opinion of Bloomberg Industry Group Inc., the publisher of Bloomberg Law, Bloomberg Tax, and Bloomberg Government, or its owners.

Author Information

Leigh Argentieri Coogan is the co-director of externships at the University of Pittsburgh School of Law and will also be a visiting professor of legal writing beginning July.

Interested in writing? Review our author guidelines, and submit pitches to Insights@bloombergindustry.com.