At the University College Birmingham Data & AI Conference, our CEO Alex Whittles delivered a keynote exploring a simple but critical message: AI is only as strong as the data behind it.

Following the keynote, the learning continued with a hands-on workshop led by Alex and our Senior AI Engineer Lewis Prince, focusing on large language model (LLM) security and prompt-based vulnerabilities. The interactive session explored how LLMs can be influenced through crafted prompts and why understanding these risks is essential for anyone working with AI.

Attendees took part in Lakera’s Gandalf challenge, an eight-level LLM security exercise designed to demonstrate how models can be persuaded to reveal protected information.

A university lecture on machine learning is in progress. Alex stands at the front of a well-lit classroom, presenting a slide titled "Machine Learning" with the subtitle "Recommendation engines." The slide features graphical representations of rating systems and user interfaces. Students sit in tiered rows, attentively listening, with some using laptops and notebooks. Several empty chairs line the side wall, emphasizing the academic setting focused on computer or data science.

Tags: , ,

« Previous Post
Next Post »