Artificial? Yes. Intelligent? Not So Much. 0
At Psychology Today Blogs, Ian MacRae suggests that observing how AI Chatbots and LLM’s deal with humor can tell us a lot about just how “intelligent” they are. He notes that “(a)Apparently, large language models have a Rickrolling problem,” and goes on to make four main points (emphasis added):
- LLMs struggle to fully grasp the deeply embedded trends and language patterns of internet culture.
- Rickrolling, a classic internet prank, reveals the limitations of machine intelligence in understanding.
- Machine intelligence mimics human communication but lacks true understanding, especially in humor.
- We don’t yet fully understand the impact of training machine intelligence on vast datasets of online language.
Follow the link for his exploration of these issues.
