AI agents, especially the ARTEMIS model developed at Stanford University, are proving to be better and more cost-effective than human hackers. In a groundbreaking experiment, ARTEMIS was deployed on 8,000 devices within the Stanford network for 16 hours, identifying and exploiting security vulnerabilities with an impressive 80% accuracy. This AI agent outperformed nine out of ten skilled human hackers, highlighting significant weaknesses that humans overlooked. The cost-effectiveness of ARTEMIS, at approximately $18 per hour compared to human hackers earning over ₹1.10 crore annually, reveals the potential for AI to revolutionize cybersecurity. Major tech companies like Google and Microsoft are already leveraging AI to combat cyber threats, signaling a shift in defensive strategies. However, experts warn of the inherent risks associated with AI’s capabilities, urging cybersecurity firms to adopt these technologies cautiously to mitigate potential errors. The implications of AI in hacking are profound and require heightened awareness.
Source link
Share
Read more