Alt Legal and Corsearch LiveAI on the Prize: Ethical considerations for emerging technologies
Alt Legal Team | January 24, 2024
On March 19, 2024 at Alt Legal & Corsearch Live, former USPTO OED attorney Emil Ali presented the session, “AI on the Prize: Ethical considerations for emerging technologies.” During this session, he explained the ethical issues to consider when using AI technology, provided real-world examples of ethical concerns stemming from the use of emerging technologies, discussed existing and proposed legal frameworks that govern the use of AI in trademark law, and shared best practices for evaluating and utilizing emerging technologies in your legal practice.
Download the presentation materials here.
Top 10 Takeaways
- Keep in mind that about 10-15 years ago, cloud-based tech solutions were not permitted. This is analogous to practice right now when it comes to AI. AI is a great tool, but we must understand the technology and how a regulator understands the technology, whether it’s your state bar or the USPTO Office of Enrollment and Discipline (OED).
- In the Crabill case, the attorney found cases on ChatGPT, didn’t read them, and didn’t verify that they were accurate. He discovered the cases were incorrect/fictitious before the motion was heard but didn’t alert the court. The judge said he couldn’t find the cases, and the attorney responded that the legal intern had found them. The attorney then filed an affidavit saying he used ChatGPT to find the cases. It wasn’t using ChatGPT that got him in trouble; it was that he lied about it and didn’t verify. With any tool you must verify—the same goes with Westlaw and Shepardizing.
- The Rule of Competence and AI: When using AI, it’s very important to understand the technology, and not necessarily the nitty gritty technical details, but specifically you need to know how it does and doesn’t work. You must also keep up with the terms of service and updates to the terms of service. Always vet the AI services you are using to confirm their data security practices and how prompts and information may be saved or used.
- The Rule of Confidentiality and AI: A concern with confidentiality and AI is if you perform a search and if you submit prompts to ChatGPT, these may become public and could create issues with confidentiality. Consider how to navigate this framework, balancing your duties to your client and your desire to do things efficiently by using technology.
- Unauthorized Practice of Law and AI: It’s improper to practice law in jurisdictions where you are not admitted, and you cannot assist others in other jurisdictions to practice law. You aren’t allowing AI to engage in law practice, but you do have a duty to supervise. After Avianca, there are more disclaimers on ChatGPT explaining it is not a lawyer. AI can be used to support lawyers but not by non-lawyers to sell legal services, nor can ChatGPT sell legal services.
- Supervision and AI: The use of AI by lawyers is similar to how we interact with non-lawyers who assist lawyers. How can you supervise the people you work with and the tools you work with? If you are the owner of a law firm, you have a distinct responsibility over the people you work with. You need to ensure you have measures in place to supervise them. Come up with rules that permit others in your office to use AI under certain circumstances, or develop rules that prohibit AI. Even if your firm says, “Our firm does not use AI,” you need to ensure that everyone in your firm understands that AI is not being used.
- California Guidelines are very helpful. The Guidelines state that you can use AI but you must supervise and review. A lawyer should take steps to avoid over-reliance on generative AI to such a degree that it hinders critical attorney analysis fostered by traditional research and writing. When a paralegal drafts an office action response, you must review and approve it before you submit it. Similarly, you must review and approve anything that comes out of AI before you submit it.
- Florida Ethics Opinion 24-1 discusses confidentiality. You do not discuss a matter without client’s informed consent. There are exceptions to confidentiality—a lawyer may reveal confidential information when it’s impliedly authorized. Is use of AI impliedly authorized? Probably not. If you think your firm will be using AI, put something in your engagement letter to disclose that you will be using AI to perform certain tasks on a private LLM and that a lawyer will always review the product.
- Florida Ethics Opinion 24-1 also makes sure that clients aren’t harmed by use of AI in terms of what they are paying. Florida wants to ensure that you do not charge 10 hours for a task that took much less time to perform using AI. If you plan to use AI, your best bet is to go to a flat fee model and inform clients that you use AI to help supplement your work by creating a first draft of the work.
- The USPTO is trying to be ahead of the curve when it comes to AI. Case law has shown that AI cannot be an inventor. Four District Courts require you to certify that you are not using AI, and if you are, you must certify that you reviewed the contents. While the USPTO hasn’t issued similar guidance, which may not be good on a public-facing system, it could require this in the future.
Speaker
Emil J. Ali
, Partner, McCabe & Ali, LLP
Emil J. Ali is a partner at McCabe & Ali, LLP where he focuses his practice on helping lawyers understand their obligations under state and federal law. As a registered patent attorney, a significant focus of his practice involves advising lawyers and law firms on all aspects of the intersection of IP and ethics matters. Emil’s work includes counseling clients on lateral transitions, malpractice avoidance, expert opinion and testimony, and respondent’s defense work before various bars and courts. You can view his musings on IP ethics issues at www.ipethicslaw.com.