We have spent weeks reading about the supposed wonders that ChatGPT does and how useful it is being for many, so much so that Open Ai wants to do Paid ChatGPT. One of its most common uses is to quickly make a text that brings together information that has been fed into it, and for this reason it is feared its massive use for school and academic work, as has just happened, which would have earned a degree in law, or at least passed certain exams.
This chatbot powered by millions of data from the Internet, including many data on US laws, has passed the exams of a US law school. In that exam he went on to write essays on topics ranging from constitutional law to taxation and torts.
Can ChatGPT do the law exams? because you have approved
Exam results with ChatGPT have been so good that educators have warned that could lead to widespread cheating. They also add that its use could mean the end of traditional classroom teaching methods, normally based on memorizing data and how to put it on paper. Although this is a debate that has been going on for years, and the role of ChatGPT would be that of an element that collects a lot of data that has been fed to it.
Jonathan Choi, a professor at the University of Minnesota Law School, put ChatGPT through the same exam that students face, consisting of 95 multiple-choice questions and 12 essay questions. In a white paper titled “ChatGPT goes to law school» Published Monday, Choi and his co-authors reported that the bot received an overall grade of C+, putting it close to an 80% approval rating.
Although this was enough to pass, ranked at the bottom of the class in most subjects. He also failed a lot on the math related multiple choice questions. They state that when writing essays, ChatGPT showed a good command of basic legal rules and a strong organization and composition. But his biggest problem is that he often had trouble spotting problems when asked open-ended questions, which is an essential skill in law school exams.
We already warned about it in an article we wrote about the chatbot, which ChatGPT makes mistakes, and only knows what is fed. It is not an infallible tool for newsrooms or anything that is not repetitive. What's more, the formal and repetitive wording resulting from its operation is what most reveals to proofreaders that what was being read was written by a bot.
