AI

Artificial Intelligence (AI) has come a long way in recent years, but as impressive as it is, it has many limitations.

AI cannot and does not replace an attorney, for one.

AI only responds to the input received. One big issue is that the way a question is asked can affect the answer given. If the question isn’t the right question asked in the right way, then the answer cannot be right, it will be biased. A good attorney removes the bias; AI cannot.

It’s easy to see how this could work for medical issues. A person might put in a list of symptoms and the AI chatbot replies with a list of possible problems and diagnoses, not taking into account the person’s medical history or other pertinent information. He or she may believe the root cause is a terrible disease, when in reality it’s an AI-generated error.

Just as AI’s powers can be misapplied in medicine, the same can happen in law.

Real-life examples. A man contacted our office concerned that his business could be found liable for wrongful termination, which could have serious consequences. An ex-employee who was not happy about being fired asked an AI chatbot about his situation, and the AI led the ex-employee to believe his firing was a “retaliatory discharge.”

The ex-employee had the AI chatbot write a letter to the employer. But it became clear that the firing was not a “retaliatory discharge,” which is a particular type of wrongful termination under South Carolina law. The AI had made several assumptions that were not true and had based its arguments on those assumptions. The employer who contacted us was relieved to find out the letter was baseless.

A similar thing happened with another employee of a different business, when a young woman upset over being fired had AI write a four-page letter threatening legal action. The letter looked great, it sounded great, but it was ersatz – four pages of legal arguments and threats built on faulty assumptions and misapplication of the law.

How could this happen? In both cases, the employees didn’t know what they didn’t know. They didn’t know enough to even ask the AI the right question about their situations, so they got wrong answers.  We don’t know exactly what prompts were used, but if one had put in something like, “I got fired from my job and I didn’t do anything wrong, is there anything I can do?” an AI chatbot could easily go from there and come up with reasons why the firing was improper and the employee had legal recourse. But it was made up.

The takeaway: Do not rely on AI to give good legal advice. It is not at the point where it can provide legal advice that’s accurate, pertinent, and unbiased.

However, you don’t have to avoid it completely. AI can be a very helpful tool when starting the research process. Just be mindful of the way you phrase questions and input you provide, as that will directly affect the answers you get back. Always verify the AI’s answers with reliable sources, too.

Call Attorney Gem McDowell

For human help with business law or estate planning in South Carolina, call Gem and his team at the Gem McDowell Law Group with offices in Myrtle Beach and Mt. Pleasant, SC, today at (843) 284-1021.