Hands-On in NYC: Testing Tesla's FSD with the Grok Chatbot

Recently, a CNBC reporter joined Tesla Model Y owner and attorney Mike Nelson for a hands-on test in New York City to experience how xAI's Grok chatbot performs while driving. Nelson, who has been using Grok for months, typically activates voice interactions with the chatbot while the car is in Tesla's "Full Self-Driving (Supervised)" mode. While crossing the George Washington Bridge, Nelson admitted he was completely disengaged from the driving process, focusing entirely on his conversation with Grok. For example, he asked Grok to adjust the seat and air conditioning, but those features are not yet fully integrated.

The test showed that Grok can follow navigation commands and answer a wide range of questions, adding convenience to the driving experience—but it also highlighted a serious safety concern. Experts note that when AI chatbots are freely used in partially autonomous driving systems, they can easily divert the driver's attention and create a risk of over-reliance. While Tesla's FSD requires the driver to remain actively supervising at all times, real-world usage may see drivers letting their guard down while interacting with AI, which is especially dangerous in busy New York traffic.

This test reveals a fundamental tension in in-car AI design: the gap between technological capability and safe deployment. Grok can explain the car's autonomous driving decisions, helping address the "black box" problem, but unrestricted conversational capabilities on the road may create new forms of distraction. Going forward, in-car AI may need to develop context-aware systems that automatically limit conversation scope when FSD is active, rather than offering full-function interaction. Safety must always remain the first principle in smart cockpit evolution.