Overview
Discover techniques for data bias, image perturbation, and jailbreaking to counter AI systems in this accessible talk about emerging defense mechanics against artificial intelligence.
Syllabus
-
- Introduction to AI and Its Potential Threats
-- Overview of AI Capabilities and Limitations
-- Understanding AI Biases and Vulnerabilities
- Understanding Data Bias
-- Types of Bias in AI Systems
-- Techniques to Identify and Mitigate Data Bias
-- Case Studies on Data Bias Impact
- Image Perturbation Techniques
-- Introduction to Adversarial Attacks
-- Crafting Perturbations to Mislead Image Recognition
-- Tools and Frameworks for Testing Image Robustness
- Jailbreaking AI Systems
-- Overview of System Jailbreaking
-- Techniques to Manipulate AI Decision-Making
-- Ethical Considerations and Legal Implications
- Emerging Defense Mechanics Against AI
-- Developing AI Robustness and Resilience
-- Role of Transparency and Explainability in Countering AI
-- Techniques for Safeguarding Against Malicious AI
- Practical Workshops
-- Hands-On Exercises on Data Bias Analysis
-- Creating and Testing Image Perturbations
-- Simulated Exercises on AI Jailbreaking
- Ethical and Social Implications
-- Long-term Implications of AI Misuse
-- Strategies for Responsible AI Development and Use
- Conclusion and Future Directions
-- Recap of Key Learnings
-- Discussion on the Future of AI Safety and Security
- Resources and Further Reading
-- Recommended Books and Articles
-- Online Courses and Tutorials for Deeper Exploration
Taught by
Tags
On-Demand
