

30 - AI Security with Jeffrey Ladish
25 snips Apr 30, 2024
AI security expert Jeffrey Ladish discusses the robustness of safety training in AI models, dangers of open LLMs, securing against attackers, and the state of computer security. They explore undoing safety filters, AI phishing, and making AI more legible. Topics include securing model weights, defending against AI exfiltration, and red lines in AI development.
Chapters
Transcript
Episode notes
1 2 3 4 5 6 7 8 9 10 11 12 13
Introduction
00:00 • 2min
Exploring AI Safety Fine-Tuning and Model Development
01:58 • 14min
Navigating AI Security Risks
15:40 • 12min
The Dark Side of AI: Exploiting Vulnerabilities for Malicious Activities
27:36 • 19min
The Battle Between AI Security and Potential Threats
46:09 • 22min
Navigating the Complexities of Cybersecurity and AI
01:08:11 • 22min
Enhancing AI Security Practices
01:29:43 • 8min
Exploring AI Models' Capabilities in Task Decomposition and Deception
01:38:03 • 3min
Exploring the Potential of AI Models with Enhanced Capabilities
01:41:00 • 4min
The Rise of Language Models in Phishing Attacks
01:44:38 • 7min
Navigating Identity Verification in the Era of AI Deception
01:52:05 • 15min
Exploring Causal Scrubbing and Network Evaluation in AI Research
02:07:16 • 2min
Exploring AI Model Capabilities and the Need for Transparency
02:09:04 • 6min