Teaching ChatGPT-4 to Break mbedTLS AES with Side-Channel Attacks

Explore a groundbreaking Black Hat conference talk that demonstrates how ChatGPT-4 was trained to perform Correlation Power Attacks on real hardware to break mbedTLS AES encryption on STM32F3 devices. Delve into the hardware and software requirements, as well as the strengths and weaknesses of the GPT-4 model for this task. Discover how prompt engineering techniques enabled GPT-4 to generate leakage models and critical parts of DUT and driver code autonomously. Gain insights into the potential implications of AI-powered side-channel attacks on cryptographic systems.

How We Taught ChatGPT-4 to Break mbedTLS AES With Side-Channel Attacks