Norton launches beta of AI Agent Protection in Norton 360

Norton today launched a beta version of its new AI Agent Protection feature within Norton 360, designed to help users safely interact with autonomous AI tools by monitoring their actions in real-time.

Checkpoint system for AI security

AI Agent Protection introduces a "checkpoint" system where actions taken by AI agents are reviewed before execution:

• Safe actions: Allowed to proceed automatically
• Confirmed threats: Automatically blocked
• Suspicious behavior: Paused for user review

Addressing new AI risks

The system aims to address the risks associated with AI agents having significant access to personal devices and data. A compromised or misconfigured agent could potentially:

• Execute harmful commands
• Expose sensitive information
• Make unauthorized changes

Compatibility and availability

The beta feature is currently available on:

• Windows systems (available now)
• Mac compatibility expected to follow

Supported developer tools include:

• Claude Code
• Cursor
• OpenClaw

Shift in security focus

This initiative reflects a shifting focus in security, moving beyond traditional malware protection to monitoring the behavior of automated systems that can act without constant supervision.

Norton recognizes that AI agents represent a new category of security concerns requiring specialized protection mechanisms.