Guest:

• Vijay Bolina, CISO at DeepMind

Topics:

• We spend a lot of time on Artificial Intelligence (AI) safety, but what about security?
• What are some of the useful frameworks for thinking about AI security?
• What is different about securing AI vs securing another data-intensive, complex, enterprise application?
• What do we know about threat modeling for AI applications?
• What attacks against AI systems do we expect to see first in real life?
• What issues with AI security should we expect to face in 3-5 years?

Resources:

• DeepMind Learning Resources
• DEFCON AI Village and videos
• CAMLIS