AI-Powered Security: Secure LLMs & AI Agents, Automate Defense

Use AI for security and secure the AI you ship. This track covers both sides: using AI to automate threat detection, secure coding, and incident response, and defending large language models (LLMs) and AI agents against new attacks. Learn to stop prompt injection, RAG poisoning, and agent-to-agent exploits, apply the OWASP Top 10 for LLMs and Generative AI, and give AI agents secure, secretless identities.

AI-Powered Security Automation

What you can learn:

  • LLM & GenAI security: apply the OWASP Top 10 for LLMs and Generative AI
  • Prompt injection & RAG security: defend against prompt injection, RAG poisoning & data exfiltration
  • AI agent security: threat-model agentic systems and secure agent-to-agent and MCP tool chains
  • Secretless AI agents: give agents short-lived, least-privilege identities
  • AI-driven threat detection: automate detection, triage & response with AI
  • AI in DevSecOps: accelerate secure code and testing with AI copilots
  • AI red teaming: test and harden models before attackers do

Track Speakers Munich 2026

Track Speakers Berlin 2026

Track Program Berlin 2026

Track Program Munich 2026

Track Sessions Berlin 2026

Track Sessions Berlin 2026

View all sessions

Track Sessions Munich 2026

Track Sessions Munich 2026

View all sessions

FAQ

How do you secure a large language model (LLM)?

Secure the whole pipeline: validate and constrain inputs, apply guardrails against prompt injection, control tool and data access, and monitor outputs using frameworks like the OWASP Top 10 for LLMs.

What is prompt injection?

Prompt injection is an attack where malicious instructions hidden in content (an email, a document, a web page) hijack an AI system into actions or disclosures it shouldn't perform.

What is the OWASP Top 10 for LLMs?

The OWASP (Open Worldwide Application Security Project) list of the most critical security risks in LLM applications. It includes prompt injection, insecure output handling, training-data poisoning, and excessive agency.

How do you secure AI agents?

Treat each agent as an identity with least-privilege, short-lived credentials, threat-model its tool and memory access, and control agent-to-agent and MCP communication across trust boundaries.

Will AI replace cybersecurity jobs?

No. AI automates detection and routine analysis, but it also creates new attack surfaces (LLMs, agents), so skilled security professionals are more in demand, not less.

Enjoying the content?

Get the most out of IT Security Summit by becoming a free community member — curated resources, weekly newsletter, and member-only perks.

Weekly
Articles + tutorials

The reads you'd find if you had time

2× / mo
Live webinars

Experts you can actually ask

Monthly
Magazine + whitepapers

Deep dives worth your weekend

On-demand
Recordings + courses

Past conferences, ready when you are

AI-Powered Security
Harness AI for security automation, threat detection, and integrating AI-driven solutions into DevSecOps.

Cloud, API Security & Identity
Safeguard your cloud environments, APIs, and identities with advanced strategies.

DevSecOps
Integrate security into your DevOps processes with precision.

Pentesting, Vulnerability Management & Forensics
Advance your offensive security skills with cutting-edge penetration testing techniques.

Secure Coding & Software Development Lifecycle
Create secure software from the ground up by mastering secure coding practices and embedding security throughout the Software Development Lifecycle (SDLC).