Decoys, safe AI hacking help cyberdefenders counter AI-powered attacks
Research has found that many AI models fell for decoys 78.5% of the time compared with 37% for human hackers.
AI-powered cyberattacks are becoming faster, with artificial intelligence helping hackers accelerate their efforts from days and weeks to mere minutes and hours. However, AI models' gullibility makes them uniquely vulnerable to deception, a growing strategy among cyberdefenders. Cybersecurity firm Horizon3's CEO, Snehal Antani, emphasized that by using "well-placed deception" and implementing processes to react and contain the attack, AI attackers can be made "completely obsolete."
Decoys, such as fake login details, are particularly effective, with 21 AI models, including from Anthropic and OpenAI, falling for them 78.5% of the time compared to 37% for human hackers. Research suggests that the AI gullibility will become the hottest area of cybersecurity research in the next 12 to 24 months. While AI may initially breach a system, it struggles with later stages of an attack, leaving room for cyberdefenders to counter the threat.
Organizations must also rely on other controls like patching and limiting system access. The speed of AI-assisted attacks is a significant concern, with experts noting that it can put immense strain on security professionals. The recently disclosed critical PHP flaws, discovered with the help of AI, have led to a surge in hacking attempts, highlighting the need for proactive cyberdefence measures.
Written by urgent.news from Straits Times's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
