Intelligent by design: Can cybersecurity cope with AI?
summary
Claims that an attacker used machine learning to blend into a corporate network in India as the first AI attack and that Darktrace’s AI outpaced the offensive algorithm lead to concrete examples: detection of an audio-only compromise of a boardroom video endpoint, discovery of unauthorized Bitcoin mining on bank servers, bribed uploads to fingerprint biometrics, and a data scientist exfiltrating IP to an AWS instance. The system inspects raw traffic, flags anomalies across 20 to 30 percent unknown IoT devices, uses a digital antibody for autonomous response, and emphasizes no training data self-learning.
AI-written summary · about
chapters (8)
- 0:00Trust attack
- 0:33First AI attack
- 1:45IoT attacks
- 3:51Machine learning
- 7:21Video conferencing example
- 9:31Bitcoin mining example
- 12:11Data exfiltration
- 14:48Conclusion
related talks
44mAI Meets Cybersecurity: Boosting Speed Without Losing Control
11mAI-Powered Cybersecurity in the Age of Advanced Threats
23mSecuring the future of work: How cyber AI learns on the job
49mGenAI and Cybersecurity: Safeguarding the Digital Age
27m"Cybertelligence": How AI is changing the face of cybersecurity
40mAI As The New Hacker Frontier
20mCatch Attackers Before They Strike with AI-powered Threat Intelligence
22mApplying AI for Threat Detection in the World of Cybersecurity