CXO Panel: Cybersecurity: Is Machine Learning The Answer?
summary
Panel argues machine learning finds patterns but needs large, high‑quality datasets and years of training to cut false positives, with supervised models limited and unsupervised learning aimed at autonomy. Demonstrated risks include adversarial attacks (models fooled by crafted inputs, e.g. a nonface image), poisoned training data, and ML‑driven phishing bots. Useful, deployed use cases are malware detection, automated data classification and on‑device transfer learning for privacy in clinical trials, while regulation and data localization create legal and cloud‑trust hurdles.
AI-written summary · about
chapters (7)
- 0:00Start
- 0:40Panelists
- 2:31Perspective of Machine Learning and Artificial Intelligences
- 17:03Why Do Babies Stop Crying
- 30:43Adversarial Learning
- 37:41Transfer Learning
- 45:53Why We Have Data Localization Laws
related talks
44mAdvancing Cybersecurity with AI and Machine Learning: Opportunities and Challenges- 44mAdvancing Cybersecurity with AI and Machine Learning: Opportunities and Challenges
16mMachine Learning--Cybersecurity Boon or Boondoggle
32mAI/Machine Learning: What is Actually Working in Cybersecurity?
45mWhat Security Measures Are Necessary to Trust in AI?
45mHardening AI/ML Systems - The Next Frontier of Cybersecurity
49mThe Intersection of AI and Cybersecurity
27m"Cybertelligence": How AI is changing the face of cybersecurity