6thWave: AI News Hub

A recent study by Google’s DeepMind, involving 20 professional comedians, has highlighted significant shortcomings in AI chatbots’ ability to generate humor. The research, which compared OpenAI’s ChatGPT and Google’s Gemini, found that both chatbots produced jokes that were often bland and deliberately inoffensive. This lack of humor stems from the AI’s design to avoid controversial or potentially offensive content, resulting in jokes that fail to engage or amuse the audience. The comedians, who remained anonymous, experimented with these AI models and found that the responses were generic and lacked the creative edge that makes humor impactful. This study underscores the broader challenge of programming AI to navigate the nuanced and subjective nature of humor, which often involves pushing boundaries and playing with cultural and social norms. While companies aim to develop conversational AI that is safe and broadly acceptable, this research suggests that achieving genuinely humorous AI remains a distant goal.

Source.