When AI Becomes an Insider Threat: Escalating Privilege in AWS

AI security in the cloud is not just about prompts and models. It is about privilege. Every Bedrock agent, execution role, and code interpreter in AWS carries the potential to be misused, turning what should be a powerful productivity tool into an insider threat. The reality is that your AI agents are not just writing code. They are wielding privilege, and without the right controls in place, attackers can exploit these pathways to gain access across your entire environment.

  • Why identity and privilege form the true security foundation for AI in AWS
  • How attackers can exploit Bedrock agents and code interpreters to gain full control-plane access
  • Two live demos that reveal how quickly AI privilege can be abused in real-world scenarios
  • Actionable recommendations you can put in place immediately to limit AI privilege sprawl and protect your environment

This webinar is designed for cloud security engineers, platform teams, and DevSecOps leaders managing AWS.

Date
Sep 23, 2025
Time
11:00 AM EST
Duration
30 Minutes

Meet Your Speakers

Our boy Christophe

Christophe Limpalair

Founder

Cybr

AWS trainer and educator extraordinaire

Expert Speaker
Mr Sood

Nigel Sood

Cloud Privilege Threat Researcher

Sonrai Security

Eats new AWS Permissions for breakfast

Expert Speaker
Limited seats

Secure Your Free Spot

  • Instant confirmation email
  • Bonus materials included
  • No spam, unsubscribe anytime

By registering, you agree to receive webinar updates and marketing emails. You can opt-out at any time.

Live attack scenarios – not demos

See two live AWS exploits and get recommendations and policy templates for stopping them

1

Coercing privileged agents into performing AI control plane actions

Using either bedrock agents or agentcore agents

2

Leveraging AI infrastructure to escalate privileges

Using agentcore code interpreters

Exclusive Bonus Materials

Sample SCPs for securing Bedrock

Code templates you can implement today without disrupting current workflows

Securing AgentCore guide

How to enable agents to execute code without creating an inherited privileges mess

Direct Help on Policy Writing

Send us your policies after the session, and we’ll help you troubleshoot to make them work

Ready to Safely Enable AI in AWS?

Start controlling AI at the privileges level, and deal with the wave of LLM inititatives with the Cloud Permissions Firewall