China Hits the Brakes on Autonomous Hype: New MIIT Rules Target Smart Driving Features

• China bans public beta testing of autonomous features and limits Level 3 capabilities.

• MIIT mandates stronger driver monitoring and transparency in feature marketing.

• Over-the-air updates for smart driving functions now face stricter approval.

---

China Tightens the Reins on Autonomous Driving: What the MIIT Crackdown Really Means

China has just dropped the mic on its domestic autonomous driving industry—by tightening the very screws it helped loosen. In a move that’s as much about consumer protection as it is about reclaiming regulatory control, the Ministry of Industry and Information Technology (MIIT) is rolling out a new set of rules that directly impact how smart driving systems are developed, tested, and marketed in the country.

Let’s break down what’s going on—and why it matters.

---

Public Testing of Autonomous Features? Not Anymore.

The most immediate and headline-worthy change is this: public beta testing of advanced driver-assistance features is now banned.

This means no more early access for flashy “city NOA” or valet-style summoning functions that haven't gone through full certification. Any smart driving system—especially those claiming high levels of autonomy—must be fully verified before hitting public roads. This is a sharp departure from the Silicon Valley-inspired "test it live" model that several Chinese automakers had been flirting with.

In other words, the MIIT just outlawed FSD cosplay.

---

Smart Driving: More Rules, Less Hype

The new guidelines require automakers to:

• Clearly define the scope and limitations of assisted driving functions

• Display feature status in real-time to the driver (no more ambiguous “Auto” lights)

• Deploy effective driver monitoring systems—especially if Level 2+ or 3 functions are enabled

• Submit OTA (Over-the-Air) upgrades for approval before pushing updates that modify any smart driving capability

These requirements are particularly tough on features like remote parking assist, one-touch car summoning, and other semi-autonomous party tricks that haven’t always worked as advertised.

What Triggered This Regulatory Shift?

The timing is not a coincidence.

Recent headlines—most notably involving the Xiaomi SU7, which made waves for its high-tech positioning but raised eyebrows for allowing “city NOA” in beta—likely forced the MIIT's hand. It was only a matter of time before regulators moved to prevent a potential safety or PR disaster.

---

Level 3 Isn’t Banned—But It’s on a Tight Leash

There’s nuance here. The MIIT hasn’t banned Level 3 autonomous driving per se. Instead, it’s set the bar much higher. Level 3 can still exist in theory, but any deployment must demonstrate system redundancy, real-time driver fallback capabilities, and meet detailed technical validation requirements before going live.

Translation? Don’t expect widespread L3 anytime soon.

Feature/Policy	Old Practice	New Regulation by MIIT
Public Beta Testing	Allowed (even on public roads)	Banned unless fully validated
Level 3 Deployment	Light oversight	Must meet high redundancy and fallback reqs
OTA Updates for Smart Driving	Freely pushed post-launch	Approval required before rollout
Feature Marketing	Often vague or misleading	Must clearly state capabilities & limits
Driver Monitoring Systems (DMS)	Optional or inconsistent	Mandatory for L2+/L3 systems

A Blow to Tech-Led Car Startups?

Make no mistake: this is a wake-up call for China’s new wave of EV brands.

Startups like Xpeng, Li Auto, and even Xiaomi, which have leaned heavily into semi-autonomous bragging rights, will need to recalibrate their playbooks. Gone are the days when marketing teams could slap a beta tag on a feature and roll it out to thousands of users in real-time traffic.

Meanwhile, traditional players like BYD and Geely—with their more conservative approach to autonomy—may actually benefit. This move levels the field a bit, favoring those who’ve taken their time with validation over hype.

---