The Security Problem No One Is Solving for AI Agents - Arcjet

In this episode of CyberBytes, David Mytton, CEO & Co-Founder of Arcjet, explains why AI agents are creating a new security problem as they move beyond answering questions to taking real actions.

David explores how security teams can control what AI agents access, the tools they use and the actions they take, while protecting against threats like prompt injection and data leakage.

We covered:

• The emerging security risks of AI agents

• Autonomous actions, prompt injection and guardrails

• Securing AI agents in real-world environments

• The rise of AI-assisted attacks

• Why cybersecurity could face major consolidation

• Building ArcJet with a lean team

• David's journey from founder to cybersecurity entrepreneur

• What's next for ArcJet and AI runtime security

As AI agents become more autonomous, securing what they can do may become just as important as securing what they can access.

Aspiron Group - https://www.linkedin.com/company/aspiron-group/

CyberBytes - https://www.linkedin.com/company/cyberbytes-as/

Joseph Cooper - https://www.linkedin.com/in/cybercoops/

Arcjet - https://www.linkedin.com/company/arcjet/

David Mytton - https://www.linkedin.com/in/davidmytton/