LLM Security in the Enterprise

Download to access this webinar replay on LLM (Large Language Model) Security in the Enterprise with industry experts from Rasa and Lakera.

In this webinar, specifically designed for business decision-makers, we explored LLMs and their impact on enterprise security, providing valuable insights into their security landscape.

The session showcased potential risks and vulnerabilities associated with LLMs, while also presenting strategies to effectively utilize them in a secure manner.

Key topics and questions that were addressed:

• Learning about Lakera and the Gandalf Experiment: Discover the motivations behind the Gandalf Experiment and gain valuable insights into its findings.

  • Lakera’s Gandalf was an experiment that went viral on social media inviting users to try and trick an LLM-based wizard to share a password.

• Understanding Prompt Injection: Explore the concept of prompt injection and learn about the most ingenious prompt injections observed to date.

• Overcoming Risks of Prompt Injection: Uncover the potential risks associated with prompt injection, like data privacy and hallucinations, and explore strategies to safely deploy LLMs in production environments.

• Systematic Security Evaluation: Gain expert advice on setting up systematic security evaluation processes to ensure safety, reduce bias, and promote transparency in enterprise conversational AI environments.