Artificial Intelligence (“AI”) has taken the world by storm. It seems no matter where you turn, AI is being promoted for all kinds of uses. By way of example, the City of Milwaukee recently announced that they will be using AI to address the increasing number of potholes around the city. It has also become apparent that more and more people are seeking legal advice from generative AI tools such as OpenAI ChatGPT, Anthropic Claude, and Google Gemini. While there is nothing inherently wrong with using AI in the legal context, there are still a number of issues one should keep in mind when using it.
1. AI is not always accurate.
AI can be a helpful tool, but it is not always accurate. A major issue currently plaguing AI is that it is well known to hallucinate. That is, AI can fabricate statutes, court cases, regulations, etc. Courts have already sanctioned Attorneys for citing to imaginary court cases hallucinated by AI. What many fail to realize is that AI can get things wrong and doesn’t guarantee a correct answer. Aside from hallucinations, it is also possible to ask AI a legal question and receive an answer that does not apply to your specific set of facts or case. For example, a Wisconsin resident may ask ChatGPT a question, fail to clarify that they live in Wisconsin, and receive an answer that would only apply to residents in Illinois.
2. AI searches are likely not protected by a privilege.
When an employee hires an attorney, the attorney-client privilege allows the client to speak openly and candidly with their attorneys about their case without fear that those discussions can be discoverable and used against them in litigation. However, when a third party gets involved in those conversations, that privilege can be waived.
A recent case out of the Southern District of New York, U.S. v. Heppner, held, among other things, that AI communications are not automatically privileged. Among the reasons given for the decision, is that these free, generative AI platforms use the information you provide (“inputs”) and the answers it gives (“outputs”) to train their models. The Court clarified that Anthropic’s own privacy policy discloses this “training” which also includes language that Anthropic reserves the right to disclose the data to other third parties. In other words, someone who signs up for a platform like Claude could not reasonably think that their communications would be completely private. However, AI’s effect on the legal field is still very much developing, and the Heppner case seems to hint that perhaps the outcome might have been different had the client been directed by his attorneys to use AI.
3. AI searches may be discoverable in litigation.
If your communications with AI are not privileged, then there is a real possibility that they can be subject to discovery requests. That is, opposing counsel can request to see any inputs and outputs generated by AI which can later be used against you. For these reasons, if you believe you have a legal issue, always contact an attorney first.
4. Contact an attorney before you use AI to research your claim.
While AI searches may seem like a simple and easy way to get information about your legal claims, it is not a substitute for skilled legal counsel. If you believe you have a legal claim against your employer, you should speak to an attorney first before utilizing AI to ensure that you are not inadvertently hurting your claims when doing so.
If you believe you have an employment-related legal claim, please contact us to set a call to discuss the specifics of your situation and how we might be able to assist.