Some studies have compared the performance of AI language models like GPT to that of human experts in various tasks. For example, a study published in the journal Nature in 2020 found that an AI language model GPT-3 performed comparably to human participants on a wide range of language tasks, including answering questions, summarizing text, and translating languages.
OpenAI’s GPT-3 and other neural nets can now write original prose with mind-boggling fluency — a development that could have profound implications for the future.
www.nytimes.com
The article notes that while AI language models have made significant advances in recent years and are capable of generating highly convincing and coherent responses, there are still limitations and risks associated with their use. For example, AI language models may perpetuate biases or inaccuracies present in the data they were trained on, and they may not always understand the nuance and context of human language.