Li Auto CEO highlights challenges in AI-driven automotive progress

By China Daily

Key Points:

• AI's Current Limitations: Li Xiang, CEO of Li Auto, argues that AI has yet to significantly reduce human workload, despite rapid improvements. He emphasizes that AI must transition from being a glorified assistant to a true productivity engine.
• Three-Stage Framework: Li outlines a framework for AI adoption:
  1. Information Tool (e.g., search engines)
  2. Assistant Tool (e.g., chatbots)
  3. Active Agent (replacing human labor in high-intensity tasks)
• Smart Driving Challenges: The auto industry's focus on intelligent driving systems faces scrutiny after a fatal Xiaomi car crash in March. China's Ministry of Industry and Information Technology has since called for clearer marketing and transparency about system limitations.
• Li Auto's AI Investments: The company has tripled its budget for AI development, focusing on a proprietary vision-language-action (VLA) model tailored for automotive data. Li compares this effort to building "the Android of smart driving."
• Open-Source Commitment: Li Auto plans to open-source its operating system, Li OS, to contribute to the community, following the example of DeepSeek's public release.
• Future Vision: Li envisions AI transforming in-car experiences, particularly for family users, by integrating high-definition vision, natural language understanding, and precise vehicle control.

Industry Context:

• McKinsey highlights the smart cockpit as a new growth area for carmakers, shifting focus from transportation to personalized in-car experiences.
• Li remains optimistic about smart driving, calling the current challenges "the darkest hour before the dawn."

Quotable:

"People assume you open-source something because it's weak. Actually, we're sharing it because we think it's strong." — Li Xiang, CEO of Li Auto