Gartner predicts that AI Agents are set to make 15% of day-to-day work decisions autonomously by 2028 – they’re here to stay and are only going to get more prolific. However, their potential is inextricably linked to a very difficult challenge – safeguarding intelligent agents against unintended outcomes, errors and ever-present cyber threats. Agents are effectively a “black box” and determining why an agent makes a particular choice is near impossible, making transparency and trust a major issue.

This talk, featuring Opencast's Director of Technology Strategy Gordon Murray alongside User Research Practice Lead Marianne O’Loughlin, highlights the challenges, risks and dependencies and explores potential ways to ensure accuracy, transparency, security and ethical standards within agentic workflows.