AI Builders Global Conference 2026. October 14-16, 2026. Online.
AI Safety & Security: Red-Teaming, Jailbreaks & LLM Hardening
This description belongs to the 2026 topic catalog. The 2027 program is being prepared separately.
As AI agents gain autonomy and interact with real systems, the enterprise attack surface expands dramatically. This session covers red-teaming methodologies for LLM applications, jailbreak prevention techniques, agent-specific security controls, kill switches for misbehaving agents, prompt injection defense, and how to build security into the entire agent lifecycle: from development to production monitoring.
Questions explored in the 2026 topic catalog
- What are the most common attack vectors for LLM applications?
- How do you build an effective red-teaming program for AI?
- What jailbreak prevention techniques are most effective?
- How do you implement kill switches for autonomous agents?
- What agent-specific security controls are essential?
- How do you monitor for adversarial attacks in production?
Who this topic is for
Security engineers, AppSec professionals, AI/ML engineers, red team operators, and CISOs responsible for securing AI-powered applications.
Related topics