Scientists from the University of Minnesota commissioned artificial intelligence ChatGPT prepare answers for final exams in four disciplines. According to the results of the study, it turned out that ChatGPT without human help is not able to achieve successful results.
This is reported by Engadget.
At the University of Minnesota, AI passed the C+ average exams. Another study was conducted at the Wharton School of Business by Professor Christian Tervish. There, for the business management exam, ChatGPT received B- and B grades .
Researchers at the University of Minnesota note that ChatGPT has done a good job of addressing basic legal norms and generalizing doctrines. However, he failed to cope with the precise definition of issues regarding cases. At the Wharton School of Business, AI perfectly solved problems with simple questions and slowed down on complex tasks. So, he could not invent the right solutions to problems in mathematics for the sixth grade.
Scientists from Minnesota suggest that it will be possible to improve the work of ChatGPT with the help of settings. Wharton researchers note that AI is able to successfully pass the exam only with the help of a person. At the same time, students can use it to improve their exam scores.
Both schools intend to prevent academic dishonesty by limiting the use of ChatGPT. In addition, it is recommended to reformulate the questions so that the AI cannot cope with them.
Recall that artificial intelligence has demonstrated its own new hypostasis. Users who installed the Replika chatbot began to complain that it was behaving inadequately.
