Why AI Agents Go Rogue and What You Need to Do About Them

Big thanks to @Cisco & @Splunkofficial for sponsoring my trip to .Conf 2026 and also for sponsoring this video.

AI agents can do things you never intended. So how do you monitor their behavior, spot security problems and decide what they should be allowed to do? I’m joined by Splunk’s Kamal Hathi at .conf26 to discuss AI agent security, observability and what an agentic SOC means for the people working in cybersecurity.

Kamal explains why setting a goal isn’t enough. You need visibility into what your agents are actually doing, clear guardrails and human oversight for important decisions. We also discuss why monitoring only a sample of agent activity can leave gaps, and how focused small language models could help reduce evaluation costs.

In this interview, we cover:
• Known agents, unregistered agents and unexpected behavior
• AI evaluations, guardrails and zero trust
• Automating SOC alert triage while keeping humans in control
• Using AI to create SOAR playbooks
• Running models locally for privacy, cost and control
• Choosing between local models and cloud models
• Whether you should still study computer science in 2027
If you’re building AI agents, working in cybersecurity or deciding what to learn next, this conversation will give you plenty to think about.

// Kamal Hathi’s’ SOCIAL //
LinkedIn: / kamal-hathi

// David’s Social //

================
Coect with me:
================
Discord: http://discord.davidbombal.com
X: https://www.x.com/davidbombal
Instagram: https://www.instagram.com/davidbombal
LinkedIn: https://www.linkedin.com/in/davidbombal
Facebook: https://www.facebook.com/davidbombal.co
TikTok: http://tiktok.com/@davidbombal
YouTube Main https://www.youtube.com/davidbombal
YouTube Tech: https://www.youtube.com/chael/UCZTIRrENWr_rjVoA7BcUE_A
YouTube Clips: https://www.youtube.com/chael/UCbY5wGxQgIiAeMdNkW5wM6Q
YouTube Emerging Technologies: https://www.youtube.com/chael/UCbY5wGxQgIiAeMdNkW5wM6Q
YouTube Shorts: https://www.youtube.com/chael/UCEyCubIF0e8MYi1jkgVepKg
Apple Podcast: https://davidbombal.wiki/applepodcast
Spotify Podcast: https://open.spotify.com/show/3f6k6gERfuriI96efWWLQQ
SoundCloud: / davidbombal

================
Support me:
================
Or, buy my CCNA course and support me:
DavidBombal.com: CCNA ($10): http://bit.ly/yt999ccna
Udemy CCNA Course: https://bit.ly/ccnafor10dollars
GNS3 CCNA Course: CCNA ($10): https://bit.ly/gns3ccna10

// MY STUFF //
https://www.amazon.com/shop/davidbombal

// SPONSORS //
Interested in sponsoring my videos? Reach out to my team here: sponsors@davidbombal.com

// MENU //
0:00 – Coming Up
0:43 – Intro
02:25 – Security Against AI
06:26 – Collecting Data on your AI
08:33 – AI Hallucinations
10:48 – Is Life Like Sci-Fi?
14:09 – Will AI Replace You?
17:16 – Not Giving Your Information to AI
18:50 – Kamal’s Prediction for the Future of AI
20:20 – Final Thoughts
21:29 – What Should you Study in 2027?
22:00 – Conclusion

Please note that links listed may be affiliate links and provide me with a small percentage/kickback should you use them to purchase any of the items listed or recommended. Thank you for supporting me and this channel!

Disclaimer: This video is for educational purposes only.
#SplunkConf26 #Splunk #Cisco

subscribe
  • David Bombal