Surprisingly easy, it turned out. In a series of experiments for a new study, Kovács found that a panel of human reviewers was unable to distinguish between reviews written by humans and those written by GPT-4, the LLM that powers the latest iteration of ChatGPT. In fact, they were more confident in the authenticity of AI-written reviews than human-written reviews.
Here is the complete storyvia Sarah Jenislawski.
