Generative AI models, including OpenAI's GPT-4, have passed a simulation of the Multistate Professional Responsibility Examination (MPRE), measuring prospective lawyers' knowledge of professional conduct rules.
AI has the potential to support ethical decision-making in the legal field, although it does not possess moral principles or the ability to distinguish right from wrong.
GPT-4 performed the best, answering 74% of MPRE questions correctly, surpassing the average human test-taker by approximately 6%. GPT-3.5, Claude 2, and PaLM 2 also achieved respectable scores. [ more ]