Explore real-world AI attacks, defenses, and privacy tactics in the age of LLMs.
As artificial intelligence rapidly evolves, so do the threats it poses—and the tools it gives us to fight back. In this talk, PhD researcher Shashvat explores how AI can both protect and compromise digital systems. You’ll learn how AIs can be “jailbroken” to leak sensitive information, how they’re being used to discover new security vulnerabilities, and why deploying AI systems that surpass human intelligence brings massive safety challenges. If you’re curious about the future of cybersecurity, AI safety, or cutting-edge tech ethics, this is a must-attend session that connects research with real-world risks.
Speaker:
Shashvat is a PhD researcher at UCL and a member of the Singapore AI Safety Hub. He works on the following questions:
What are the applications of quantum computing?
What are security risks of quantum computers and how can we mitigate the expected harms?
How will quantum computing affect AI capabilities and AI risks?
He has previously worked as a market analyst with The Quantum Insider, and won top academic prizes at Oxford for both Computer Science and Philosophy.
Agenda:
Speaker Introduction & Background
Understanding Adversarial Attacks on LLMs
Defense Strategies: Safety, Alignment, and Privacy
Real-World Tools: RAG Systems & Automation Use Cases
Wrap-Up: AI+X Programs & Global Learning Opportunities