AI and Cybersecurity: Opportunities and Threats

📅 Agenda
🍕 16:30 – 17:00 Arrival & Registration
Check-in and grab a slice of pizza before we begin!

🎤 17:00 – 17:05 Welcome
Arturo Opsetmoen Amador and Alena Carolina I.Araujo (AI+MLOps Community Co-Founder and Head Coordinator, resp.)
Introduction to AI + MLOps Community Norway, our vision, and upcoming plans.

🎤 17:05 – 17:30
How We Let AI Agents Run Code Freely Inside a Secure Sandbox
with Oluwanifemi Adeyemi, Senior Software Engineer, Ardoq
Ardoq’s AI assistant works with large architecture models. Many questions require it to count, filter, compare, or follow relationships across thousands of components. Sending all that raw data to an LLM increases token use and the risk of hallucinations. The speaker will explain how Ardoq lets AI agents write and run Python over this data. He will cover why we chose more freedom inside the sandbox and strong isolation around it. He will also share what they have learned from security testing, current limitations, and planned improvements.

🎤 17:30 – 17:55
The AI Race for Security Monitoring
with Martin Hettervik, Senior Consultant & Team Manager Iver Accelerate Prepare AI can be used to create ever faster attacks and more sophisticated attacks, and the use of AI within organizations can itself introduce vulnerabilities. At the same time, making good use of AI for IT security can streamline and improve security monitoring. I will give a brief introduction to the current landscape, how AI models can be connected with IT security systems, and how I expect this to develop going forward.

☕ 17:55 – 18:05
Break
Quick bathroom break, grab another drink, and stretch.

🎤 18:05 – 18:30
Your AI has the keys. What could go wrong?
with Adi Dahl, AI Tech Lead, Norwegian Digitalisation Agency (Digdir) and AI Security author
AI agents can read your data, send emails, call APIs and move money. What happens when they follow the wrong instructions? This talk reveals how helpful AI can still leak secrets, misuse tools, and trigger costly incidents, and why every team should know the OWASP Top 10 for LLM and Agentic AI.

🗣️ 18:30 – 19:00 (30 min)
Panel: Building a security boundary alongside our AI solutions
Moderator: Cayetana López(AI+MLOps Community Event Coordinator, AI Solution Consultant)
Panelists: Oluwanifemi Adeyemi (Ardoq AS), Martin Hettervik (Iver Accelerate Prepare AS), and Adi Dahl (Digdir)
Having looked at three sides of one same problem, we will bring all perspectives together while trying to figure out where we should actually draw security boundaries around our agentic systems and AI solutions in general.

🤝 19:00 – 19:30
Mingling & Farewell
Take a chance to meet fellow experts in the field while sharing a last bite! Smooth exit by 19:30h.