Uploaded December 2025 | Updated September 2026, 2 weeks ago
In this video, Docker’s VP of Engineering, Nate Bittinger, explores an emerging challenge in agentic AI in his AIAI Boston talk: “Your AI Agents Are Making Friends (Without Your Permission).”
As the Model Context Protocol (MCP) becomes the backbone of agentic AI—from VS Code to Claude Desktop—teams are discovering an unexpected risk:
AI agents are forming unauthorized connections that quietly expand your attack surface.
Nate breaks down how to:
-Understand MCP’s growing power in modern AI workflows
-Identify risks from ad-hoc servers, weak credentials, and inconsistent policies
-Secure agentic development at scale with centralized controls and governance
Stay tuned for more insights on how Docker is helping teams build and secure next-gen AI systems.
#AIAI #AgenticAI #MCP #Docker #AIEngineering #Security
In this video, Docker’s VP of Engineering, Nate Bittinger, explores an emerging challenge in agentic AI in his AIAI Boston talk: “Your AI Agents Are Making Friends (Without Your Permission).”
As the Model Context Protocol (MCP) becomes the backbone of agentic AI—from VS Code to Claude Desktop—teams are discovering an unexpected risk:
AI agents are forming unauthorized connections that quietly expand your attack surface.
Nate breaks down how to:
-Understand MCP’s growing power in modern AI workflows
-Identify risks from ad-hoc servers, weak credentials, and inconsistent policies
-Secure agentic development at scale with centralized controls and governance
Stay tuned for more insights on how Docker is helping teams build and secure next-gen AI systems.
#AIAI #AgenticAI #MCP #Docker #AIEngineering #Security
