Cover Image for AI Security: Real Breaches, Real Defences
Cover Image for AI Security: Real Breaches, Real Defences
Avatar for AI Safety UAE
Presented by
AI Safety UAE
4 Going

AI Security: Real Breaches, Real Defences

Google Meet
Registration
Approval Required
Your registration is subject to host approval.
Welcome! To join the event, please register below.
About Event

​✨ This is the community's recurring AI Safety UAE Meetup in Dubai! ✨

​AI Security: Real Breaches, Real Defences

​Have you heard of the HuggingFace/OpenAI incident or the Fireworks AI & Notion data breach? Or the case where nine Mexican government agencies were breached between December 2025 and February 2026 by someone posing as an authorized bug-bounty tester?

​If that startles you, this meetup is for you. Join us as we talk about AI security, the practices to follow, and what we can do about it.

​Meet our speakers

​🎤 Isham Rashik, AI Engineer at FutureOS and Community Leader/Co-founder of CodersHQ's AI/ML Community, will discuss how main vs abliterated models work for AI safety. Main models are safe for everyday use, while abliterated models are used to test a model's defences and simulate real-world breaches without the AI refusing to cooperate. He'll explain why both matter.

​🎤 Sayed Shaista, Founder & CEO at Carnelian Technologies (a Dubai-based software QA and testing company), will share the security and quality gaps that keep repeating when testing AI-built (vibe-coded) apps, and how to turn those findings into a practical pre-launch checklist for builders.

​💬 Open to anyone interested, involved, or just starting out in AI safety

​📍 Online

​🗓 Friday, November 20, 2026

​🕔 5:00 PM – 6:00 PM UAE Time.

​This is a chance to connect, share ideas, and build momentum around one of the most important challenges of our time. If that sounds like your kind of evening, we would love to see you there.

​

​Note: Seating is limited, so kindly arrive on time to secure your spot.

Avatar for AI Safety UAE
Presented by
AI Safety UAE
4 Going