Smart Locks' Facial Recognition Vulnerabilities Exposed in China

The event involves AI systems explicitly through the use of facial recognition technology in smart locks. The malfunction or inadequacy of the AI system's liveness detection and anti-spoofing features has directly led to security vulnerabilities that allow unauthorized access (harm to property and privacy). The article describes actual security incidents (successful unlocking with photos) and risks of data interception, constituting realized harms. Therefore, this qualifies as an AI Incident due to the direct link between AI system malfunction and harm.

The event involves AI systems explicitly through the facial recognition technology in smart locks. The malfunction or insufficient anti-spoofing capability of these AI systems has directly led to security vulnerabilities that can cause harm to property and personal safety. The article documents realized security risks (harm) due to the AI system's failure to prevent photo-based spoofing, which is a direct AI Incident as per the definitions. Therefore, this qualifies as an AI Incident rather than a hazard or complementary information.

The event involves AI systems as smart locks use AI technologies such as facial recognition and encrypted data transmission for authentication and remote control. The reported security flaws in AI components (facial recognition anti-spoofing failure) and data handling (unencrypted transmission) have directly led to realized harms, including unauthorized unlocking and privacy risks, which constitute harm to persons and communities. Therefore, this qualifies as an AI Incident because the development and use of AI systems in these smart locks have directly led to security harms. The article does not merely warn of potential risks but documents actual vulnerabilities and successful exploits, confirming realized harm.

The article explicitly mentions AI systems in the form of facial recognition technology used in smart locks. The vulnerabilities allow unauthorized unlocking using photos, which directly leads to harm to property and personal security. Additionally, unencrypted data transmission poses privacy risks. These constitute direct harms caused by the AI system's malfunction or design flaws. Therefore, this qualifies as an AI Incident under the framework, as the AI system's use has directly led to harm.

The event involves AI systems because facial recognition technology is an AI system used for unlocking. The article explicitly states that three smart locks with facial recognition can be easily fooled by photos, indicating a malfunction or weakness in the AI system's anti-spoofing capabilities. This has directly led to security harms (unauthorized access), which is harm to property and user privacy. Additionally, the unencrypted transmission of sensitive data increases risk of unauthorized control. Therefore, this qualifies as an AI Incident due to realized harm caused by AI system vulnerabilities in use.