Monday 23 September 2024, 11am-12noon (BST)

As large language models (LLMs) continue to advance and permeate various industries, they also introduce a growing list of cybersecurity concerns. From attacks like jailbreaks and prompt injections to more subtle risks, LLMs have various failure modes that can affect your applications.

In this presentation, we delved into the challenges of developing secure and trustworthy LLM-based solutions, explored practical strategies for mitigating risks, and examined how to align these systems with your organisation’s objectives.

Chaired by

Tarquin Folliss OBE (info), Vice Chairman, The SASIG

Presented by

Vladislav Tushkanov (info), Group Manager of Kaspersky MLTech, Kaspersky