In a new study designed to test the co-creative capabilities of large language models (LLMs), internet memes created by OpenAI's GPT-4o model were, on average, rated funnier, more creative and more ...