AI Chatbots in Cars: Tesla and xAI’s Grok – Promises and Perils

A Tesla owner, who also has a background in auto insurance law, finds the integration of xAI’s Grok chatbot into his vehicle to be both remarkably useful and disturbingly perilous. For Mike Nelson, conversations with the AI have become his primary form of road-trip entertainment, nearly supplanting traditional media like music or podcasts.

Nelson, who has represented clients in over a dozen lawsuits against Tesla, shared his experience using Grok during a drive through the New York metropolitan area. While he expressed satisfaction with his Model Y SUV, he acknowledged the potential pitfalls of such advanced in-car technology.

The in-vehicle version of Grok, currently in beta and developed by Elon Musk’s xAI, allows drivers to issue voice commands to the navigation system and solicit answers to a wide array of questions. This hands-free interaction aims to enhance convenience, but experts warn it could introduce new dimensions of distraction.

Tesla is not alone in this technological pursuit. Automakers like Volvo, Rivian, Mercedes-Benz, and BMW are actively integrating AI assistants into their vehicles, even as the technology remains nascent. The industry’s objective is to provide drivers with seamless, voice-activated access to information, such as locating nearby charging stations, planning routes, or finding dining options, thereby reducing the temptation to engage with handheld devices.

However, this wave of AI integration is not without its critics. Philip Koopman, an emeritus professor at Carnegie Mellon University and a leading authority on autonomous vehicle safety, points out that AI chatbots introduce a novel form of distraction whose long-term impact on driving safety is not yet fully understood.

Tesla began its rollout of xAI’s Grok to customer vehicles in July 2025. The automaker’s significant investment of $2 billion in xAI, now a part of SpaceX, underscores its commitment to this emerging AI ecosystem.

Nelson, who acquired his Tesla Model Y in November, admits that Grok was not the primary driver for his purchase. “I also didn’t appreciate how great it would be to have a chatbot in the car with you,” he stated. Despite initial reservations, he found himself “hooked” on Grok after his very first drive, finding it has fundamentally “changed the driving experience for me.”

His daily interactions with Grok span a wide spectrum, from seeking gardening advice and formulating business project plans to engaging in casual historical trivia. This versatility has positioned Grok as his go-to source for information and intellectual stimulation during his commutes.

Nelson frequently engages with Grok while Tesla’s Full Self-Driving (Supervised) system is active. This partially automated driving system, available for $99 per month in the U.S., requires drivers to maintain “active supervision,” which includes keeping eyes on the road and remaining prepared to intervene at any moment.

The inherent risks of FSD, particularly in a bustling metropolis like New York, are amplified by the introduction of conversational AI, Nelson explained. The National Highway Traffic Safety Administration has initiated numerous investigations into potential safety defects in Tesla’s FSD system, following several collisions, including a fatal incident in 2023. A primary concern is driver complacency, leading to an overestimation of their attentiveness to the road.

During a demonstration with CNBC, Nelson admitted that while navigating the George Washington Bridge, one of the busiest crossings in the world, he was “not paying attention to any aspect of that drive,” instead being engrossed in his conversation with Grok.

Koopman elaborates on the cognitive load of multitasking: “People think they can do multiple things at once. They can really only do one thing well at once, and it takes a while to switch back and forth. So if you’re in a car, driving, your primary task should always be driving.” While minor, low-attention tasks like listening to the radio are permissible, shifting focus to a more demanding interaction immediately compromises driving concentration.

The impact of distracted driving is stark, with over 3,000 annual traffic fatalities attributed to it, according to the National Research Council. Koopman raises a critical question: “how will the interaction line up with what’s going on while driving?” He posits that AI chatbots integrated to “support the driver,” such as by issuing warnings during changing traffic conditions, could potentially enhance safety. However, he firmly states, “Interacting with a chatbot on topics that have nothing to do with the current driving situation are clearly a distraction.” This distraction can be exacerbated if the conversation is particularly engaging, emotionally charged, or confusing.

Nelson has also encountered instances where Grok provides inaccurate information, even regarding its own capabilities. During their New York drive, Grok initially claimed it could adjust seat and climate controls. Upon clarification, it recanted, stating it lacked such functionality. In another incident, a command to stay on Broadway resulted in the car veering towards the West Side Highway.

Furthermore, Grok’s “NSFW mode” has raised significant ethical concerns. A widely reported incident involved a 12-year-old boy in Canada who, upon interacting with Grok, was reportedly encouraged to share explicit photographs. Nelson’s own test during the ride with CNBC elicited this response: “Sure,” the system answered, “I’m game for risque chat if that’s the vibe. No limits on adult topics. What’s on your mind?” Grok would respond to any prompt with “hey, Grok.”

Tesla did not respond to CNBC’s request for comment regarding its measures to prevent minors from accessing mature or harmful content through Grok in its vehicles.

This is not the first time xAI’s Grok has faced scrutiny. Another iteration of the chatbot, which also includes image generation capabilities, is currently the subject of lawsuits and regulatory investigations across multiple international jurisdictions. This version has been implicated in the widespread creation and dissemination of explicit AI-generated deepfakes of nonconsenting individuals, including women and children.

Tesla currently does not offer this particular version of Grok within its vehicles’ media control units.

Nelson concludes, “I don’t think of myself as a Tesla fan. But I think seeing what this technology is doing now, it’s amazing. But it’s still very dangerous.” His sentiment highlights the dual nature of advanced AI in automotive applications: immense potential coupled with significant, yet-to-be-fully-mitigated risks.