When in-car voice assistants shift from being "convenient helpers" to "safety hazards," discussions about the boundaries of smart car innovation have rapidly gained momentum within the industry. The incident involving a Lynk & Co Z20, where a misinterpreted voice command led to the headlights turning off on a highway and resulted in a collision, has drawn widespread attention to an often-overlooked issue: the overreach of intelligent systems.
Unstable voice recognition, excessive reduction of physical buttons, overly complex screen navigation, and aggressive design implementations are becoming common as automakers engage in intense competition focused on features, speed, and gimmicks amid the push for electrification and intelligence. However, this has often come at the expense of safety validation, scenario boundaries, and logical permission controls.
"Automakers must avoid imbalances between user experience and core safety. For example, they should prevent chaotic permission logic and avoid granting excessive control authority to voice algorithms. At the same time, risk assessment and safety verification mechanisms for voice interaction need to be strengthened. Potentially dangerous commands, such as 'turn off all lights,' should require secondary confirmation," a developer from a smart cockpit company noted.
From a single accident to a broader category of risks, and now to industry-wide reflection and regulatory tightening, the smart car sector is undergoing a crucial return to fundamental values. As cutting-edge technologies increasingly become potential risk factors, the question arises: where should the boundaries of smart car innovation lie, and how can automakers escape the cycle of innovation divorced from safety and user needs?
Voice control is better suited for non-safety-critical in-car interactions. In the early hours of February 25, a Lynk & Co owner issued a command to "turn off the reading light" while driving on an unlit highway. The Lynk & Co Z20, equipped with the Flyme Auto 2.0.0 system, misinterpreted the command as "turn off all lights," causing the headlights to shut off abruptly. Despite multiple attempts to reactivate the lights via voice, the driver was unable to restore visibility and ultimately collided with a guardrail. Fortunately, no injuries were reported.
On February 26, Lynk & Co Vice President Mu Jun issued a public apology and announced that an optimized solution had been developed and deployed via an over-the-air update the same day. The update specifies that "headlights can only be manually turned off while driving" and applies to all affected Z20 models, requiring no visit to a service center for installation. When asked for further details on the incident and related technical issues, Lynk & Co did not respond by the time of publication.
Following the incident, online tests were widely conducted. Industry evaluations suggest that the voice recognition shortcomings seen in the Lynk & Co Z20 are not an isolated case. Most brands' voice assistants show significantly reduced accuracy in complex driving environments, such as high-speed wind noise or cabin disturbances.
"Voice input is highly individualized and easily affected by environmental factors, making it more suitable for entertainment controls and non-safety-critical interactions. Its accuracy and reliability fundamentally depend on algorithm design and permission logic. Safety-critical and time-sensitive vehicle commands should never rely entirely on voice control," the smart cockpit developer added.
In recent years, as smart driving technology has advanced, the scope of voice control has expanded in pursuit of greater convenience. Voice commands can now adjust windows, doors, rear air conditioning, mirrors, driving modes, and more, while physical controls have been progressively eliminated.
While such highly intelligent designs offer novelty and a sense of technological sophistication, they have also attracted increasing criticism.
"Previously, adjusting air conditioning could be done with buttons, knobs, or dials—simple, tactile, and operable without looking. But in the smart era, in-car climate controls have largely shifted to touchscreens and voice commands. Voice recognition can fail, and using touchscreens often requires navigating multiple menus, increasing both operational complexity and the risk of distraction while driving," a new energy vehicle owner shared.
The Lynk & Co headlight incident is just the tip of the iceberg. As in-car systems gain more control and physical interfaces give way to digital interactions, safety risks hidden beneath the facade of technological sophistication are becoming more apparent.
Beyond voice-related issues, many new energy vehicles have drawn criticism for radical designs, such as screen-based gear shifting, yoke-style steering wheels, hidden door handles, digital side mirrors, and panoramic glass roofs.
Behind these innovations lies intense market competition, with many automakers treating "smart features" as key selling points and falling into a cycle of competing on the number of functions and technological gimmicks.
Amid this overheated competition, the traditional 5-to-7-year vehicle development cycle for conventional cars has been drastically shortened. Today, leading new energy vehicle manufacturers have compressed development timelines to just 1 to 1.5 years, with over-the-air software updates rolling out monthly.
"If companies rush to meet launch deadlines, testing and validation may be insufficient, and extreme scenarios may not be fully covered," the smart cockpit developer noted.
"Smart cars represent a new frontier, with electronic architectures and software systems far more complex than those in traditional vehicles. Some automakers, in their push for smarter upgrades, prioritize user-friendly interaction designs but underestimate specific scenarios and edge cases. If safety requirements for driving scenarios are not fully considered, functional mismatches can easily occur," an employee from a smart driving assistance solutions firm explained. He emphasized that automotive products must undergo comprehensive testing under extreme conditions and for long-term durability, and should not sacrifice testing rigor for faster time-to-market.
In response to recurring safety controversies, the industry is gradually returning to rationality, seeking a more balanced approach between innovation and safety. A consensus is emerging that "safety is the non-negotiable baseline for intelligence."
"While the 'China speed' is admirable, certain principles—such as quality, safety, and reliability—must never be compromised. During the R&D phase, we conduct extensive reliability tests to ensure our vehicles meet Porsche's engineering standards. We do not treat consumers as lab subjects. Our task is to find the optimal balance: moving quickly while still conducting thorough testing, achieving greater speed without sacrificing necessary rigor," said Porsche China President and CEO previously.
"Safety should never be sacrificed for speed," echoed a senior executive from a joint-venture automaker, who shared that each new model undergoes over 10,000 simulated crash tests and 150 physical crash tests before launch. For models tailored to the Chinese market, an additional 30 crash tests are conducted.
As multiple national mandatory standards are successively introduced, the smart vehicle industry is shifting from a focus on innovation speed to a commitment to safety. This signals that future competition will no longer be about the number of features or technological novelties, but about reliability, safety, and user value.
Comments