“GEEKCON2025”上海站收官 业界共探AI与机器人安全等新挑战

Core Insights - The integration of artificial intelligence with the physical world has raised significant security concerns across various applications, as highlighted during the GEEKCON2025 event in Shanghai [1][4] - The event showcased vulnerabilities in smart technologies, including humanoid robots, AI smart glasses, and drones, emphasizing the need for enhanced security measures [2][3] Group 1: Security Vulnerabilities - A demonstration revealed a logic flaw in humanoid robots that allows attackers to gain code execution permissions through specific voice commands, posing systemic risks to robot clusters [1] - AI smart glasses were compromised in a challenge, allowing attackers to remotely control the camera and microphone, thus threatening user privacy [2] - Multiple vulnerabilities were identified in commercial drones, with one critical flaw enabling attackers to spoof control signals, which is crucial for the burgeoning low-altitude economy [2] Group 2: Industry Challenges - A study on mobile device biometric authentication systems uncovered design flaws that could allow attackers to bypass security measures, raising concerns about sensitive data exposure [3] - The industry faces challenges such as insufficient emphasis on cybersecurity in emerging sectors, inadequate investment in security by traditional industries, and a decline in security research positions, which weakens foundational security capabilities [3] Group 3: Initiatives and Future Directions - GEEKCON has established a quantitative security rating system for smart devices, aiming to make security comparable and perceptible, similar to health ratings for restaurants [4] - The event organizers are committed to promoting responsible disclosure and handling of security vulnerabilities, collaborating with global security researchers to safeguard digital security in the intelligent era [4]