Fireside chat
summary
Focus on practical failures and fixes: critical infrastructure AI breaks because of poor or insufficient data, incompatible vendor formats, and attacks that target training data. Concrete measures include mandatory data standards and human-labeled golden datasets, continuous monitoring and encryption, and incident response plans — the Colonial Pipeline outage halts fuel for about a week. Governments push zero trust and multinational sharing of remediation playbooks. Defenses include validating inputs, guarding against data poisoning, and using computer vision to auto-detect perimeter anomalies.
AI-written summary · about
chapters (18)
- 0:00AI in critical systems
- 0:45Critical systems and AI use cases
- 2:08Data challenges in AI
- 3:56Regulatory frameworks for AI
- 7:07Security risks of generative AI
- 9:14Deep fake and phishing attacks
- 11:01Human error in AI deployments
- 12:49Team composition for AI systems
- 15:43Data confidentiality integrity availability
- 17:53Ethical considerations in AI
- 21:06AI innovation versus regulation
- 22:33International AI security collaboration
- 23:58US AI threat preparedness
- 27:31Cyber attacks on critical systems
- 30:23Grid backup and resilience
- 34:14Data poisoning prevention
- 38:47Bias in AI training data
- 44:25Closing remarks and Q&A
related talks
22mFireside Chat - AI Security Vulnerabilities
40mDefending Operational Frontlines with Industry Experts with...
12mSecuring AI-Driven Finance
01h 33mInnovate with Resilience!
17mA Resilient Journey
18mDefending the unknown using AI and cybersecurity
42mReal-World Security Implementations and Key Lessons
33mLearning from Past Security Breaches: Strengthening AppSec Efforts and Focus