/dev/world 2026

What do Developers Need to Know About AI Attacks, Securing AI and Coding Agents

AI is an amazing tool for developers - but also - it's an amazing tool for attackers. Your coding agent reads files it's told to trust, runs commands on your behalf and touches your credentials, cloud and potentially production systems. If you YOLO this, a coding agent can be talked into misbehaving, working for the attacker instead of you. The goal of this talk is to see how developers are being targeted in the age of AI. What has already gone wrong in the real world - what the attackers are doing against developers - and how can we help mitigate against these risks.

About the presenter

Louis is one of two /dev/eryworlders that has been to every /dev/world. He's known by people as the Brain Dev Sec guy as his interests are Neuroscience, Coding and Information Security. He's currently the lead cyber instructor at Lumify Work, Cyber Training Business of the Year 2025 and 2026 (CyberDaily) and he teaches most of the well-known security certifications to companies all over Australia and New Zealand.