AI Village Interview
Chris visits the DEF CON 31 AI Village to discuss large language model red teaming, generative AI security, and algorithmic vulnerabilities.
Summary
Chris interviews an organizer at the DEF CON 31 AI Village, highlighting the massive Generative Red Team Challenge (GRT) organized in collaboration with the White House and leading artificial intelligence research labs.
[00:57] The organizer explains the scope of the historic public red teaming event, where thousands of hackers were granted access to state-of-the-art commercial large language models (LLMs) to identify flaws, algorithmic bias, prompt injections, and safety jailbreaks.
[01:38] The interview details other interactive exhibits in the AI Village, including hardware demonstrations running image classification models on custom FPGA accelerators with FermiLab engineers, diffusion model manipulation using ControlNet, and hands-on preparation stations hosted by the Defense Digital Service (DDS).
(AI summary — Gemini Flash)
Related talks
Blacks in Cybersecurity Village·33 min
Levone Campbell
Explores the dual role of artificial intelligence in scaling hyper-targeted phishing campaigns while powering real-time behavioral defenses.
Videos·52 min
Solomon Sonya
Solomon Sonya explores 🇬🇧 AI for Cybersecurity: Applying Machine Learning to Enhance Malware Analysis.
Briefings·26 min
Jiashuo Liang · Guancheng Li
Introduces a novel gradient-based prompt-injection technique that can generate universal triggers to manipulate open-source Large Language Model (LLM) outputs.