Pranav Shikarpur

Conf42: Large Language Models (LLMs) 2024

Getting AI to Do the Unexpected

The top three vulnerabilities with LLM apps currently are Prompt Injections, Insecure Output Handling, and PII data leakage. In this session, attendees will learn about these prompt hacking vulnerabilities, mitigation strategies, and the importance of 'secure by design' practices in LLM app development.