Industry Briefing
A single destination for timely, editor-curated robotics news from around the world.
Carnegie Mellon’s hacking team wins fourth straight, record ninth overall DEF CON Capture-the-Flag title
Carnegie Mellon University's Plaid Parliament of Pwning (PPP) has achieved a remarkable feat by winning the DEF CON Capture-the-Flag (CTF) competition for the fourth consecutive year, marking its ninth overall victory in the past 13 years. This prestigious event, known for its intense cybersecurity challenges, took place during the annual DEF CON conference, a leading gathering for hackers and cybersecurity professionals. The team's consistent success underscores their expertise and dedication in the field of hacking and cybersecurity, solidifying their reputation as the most successful team in the competition's history.
Uncategorized
AI Hacking Threats Extend to Humanoid Robots: Seoul Hosts 'Judgment Day' Challenge
The 2026 Seoul AI Security Forum has underscored the critical importance of physical security for artificial intelligence systems, particularly as hackers increasingly target humanoid robots. The event featured a live red team challenge known as "Judgment Day," which focused on testing the vulnerabilities of a U-Tree G1 robot. This demonstration highlighted the necessity for proactive security measures in response to the rapidly evolving landscape of AI technology. The forum serves as a vital platform for addressing the challenges posed by cyber threats to advanced robotics, emphasizing the need for robust defenses to protect these innovative systems.
AI Security Humanoid Robots Red Teaming Robustness Testing
Voice AI Systems Are Vulnerable to Hidden Audio Attacks
Researchers are set to unveil alarming findings regarding AI-powered voice and audio tools at the IEEE Symposium on Security and Privacy in San Francisco next week. The study reveals that modified audio clips, imperceptible to human ears, can manipulate large audio-language models (LALMs) to execute unauthorized commands with a success rate between 79 and 96 percent. This vulnerability allows attackers to control devices, conduct sensitive web searches, and even send emails containing user data without the user's knowledge. The research, led by Meng Chen, a Ph.D. student at Zhejiang University in China, demonstrates that these attacks can be executed in real-time and do not require the attacker to have full control over the user's instructions. Instead, adversarial audio can be embedded in various media, such as online videos or voice notes, making it a pervasive threat. The technique, dubbed AudioHijack, exploits a critical flaw in LALM design, allowing malicious instructions to be hidden within manipulated audio clips. The researchers tested their method on 13 leading open models, including those from Microsoft and Mistral, and found that their attacks could be adapted to commercial models as well. While Microsoft acknowledged the importance of the research in enhancing model resilience, Mistral did not respond to inquiries. The study highlights the challenges in defending against such attacks, as common defenses have proven largely ineffective, underscoring the urgent need for improved security measures in AI audio technologies.
Hacking Digital-audio Adversarial-attacks Open-source-software CybersecurityRobotics needs a service framework.
RSF defines a common language for robot service capability, lifecycle operations, certification pathways, and service-provider networks.
