The conversation around the use of artificial intelligence in legal examinations took an intriguing turn as the California justices approved the State Bar’s request to adjust the scores of February’s California Bar Exam. This decision, rendered on a Friday, came amid discussions about the reliability and origins of the questions administered during the exam.
A notable development in the February examination was the utilization of artificial intelligence to draft a portion of the multiple-choice questions. Specifically, it was disclosed that 29 of these questions were generated by ChatGPT. This raised eyebrows and ushered in discussions concerning the credibility and appropriateness of using AI in such high-stakes testing environments.
Despite these revelations, the justices required that the upcoming July exam adhere to more traditional forms of examination. They instructed the California Bar to apply the Multistate Bar Exam (MBE) format for the multiple-choice sections, which are typically considered a standard across various jurisdictions in the United States.
The justices remain cautious, expressing that while the February exam contained a sufficient number of dependable questions, they are vigilant regarding the selection process for these questions. They acclaimed the necessity for scrutiny, recognizing the pending audits of the February exam, which are expected to shed more light on the procedure utilized in question formulation.
As the topic of AI’s role in legal assessments progresses, this incident stands as a signal for ongoing evaluations and necessary adjustments to ensure the integrity and fairness of legal licensure examinations. For further details, visit the full article on Bloomberg Law.