Find out why a recent study conducted by German researchers found that the humorous capabilities of ChatGPT 3.5, a popular chatbot, are limited. Learn about the study's findings and what this means for the future of chatbots.
German aerospace researchers discovered that OpenAI's ChatGPT-3.5 model has limited joke knowledge, with over 90% of generated jokes repeating from a set of 25. While the AI struggled with new sequences and failing jokes, it exhibited knowledge of wordplay and humor elements to some extent.