/Introduction/

Recently, the tech community has been abuzz with Microsoft's ChatGPT. As a chatbot launched by AI company OpenAI in November 2022, it can engage in conversations by learning and understanding human language, interact based on chat context, and even perform tasks such as writing emails, video scripts, copy, translations, and code. Its intelligence far surpasses all current human-computer interaction models. The sensation caused by ChatGPT has also put unprecedented pressure on major tech companies like Google and Baidu, whose primary business is search. They have since launched their own ChatGPT-like models, but their actual performance remains questionable.

So, if text learning can reach such a level of intelligence, what can ChatGPT bring to the autonomous driving industry? Can it provide a breakthrough for the decision-making challenges that have plagued practitioners for years?

Decision-Making Challenges in Autonomous Driving

In the autonomous driving industry, decision-making and planning, as the brain of the entire system, are core algorithmic modules. They play a crucial role in the realization of autonomous driving functions and the assurance of