AI agents are coming to Rivian’s cars

Rivian vehicles will soon offer AI assistant capabilities beyond what the carmaker directly provides.

Rivian plans to integrate third-party AI agents into its vehicles, allowing drivers and passengers to interact with external AI models like those from Anthropic and OpenAI. CEO RJ Scaringe announced this at TechCrunch Disrupt, building on Rivian's existing in-car AI assistant. The company has developed its own agentic framework to understand user commands and route them to appropriate software functions, ranging from climate control to news summaries. While the current assistant can access external apps like Google Calendar, the next phase involves connecting with advanced AI models. Rivian is developing the necessary "plumbing" to ensure these agents function safely, explicitly stating that vehicle control systems like braking and steering will not be delegated to external AI. The integration is expected to enable more sophisticated contextual interactions, such as planning meals or accessing information from past conversations across devices.

Rivian's strategy to allow third-party AI agents into its vehicles, rather than solely relying on in-house development or captive AI models, signals a potential shift towards an open ecosystem for automotive AI. By building its own agentic framework, Rivian positions itself as a platform provider, enabling integration of diverse AI capabilities and potentially fostering innovation within its car software. This approach could lead to richer, more personalized in-car experiences by leveraging the specialized strengths of different AI developers.

This move by Rivian, to integrate sophisticated third-party AI agents, has broad implications for the automotive industry's adoption of AI. It challenges the trend of vertically integrated AI solutions, suggesting a future where car interiors become extensible computing environments. For consumers, it promises enhanced functionality and convenience, but for Rivian and its AI partners, it raises complex challenges around data privacy, security, and ensuring reliable, safe operation, especially when dealing with AI models that learn and evolve.