AI-Assisted Canary Token Experiment Demonstrates Cyber Deception

Summary

A livestream experiment used Codex to create and deploy canary token lures in a virtual machine, with a webhook receiving alerts. A manually scanned QR-code token triggered successfully; the project remains a proof of concept.

Key points

• Canary tokens act as deception tripwires, alerting defenders when someone interacts with a lure.
• The proposed workflow uses AI to inventory a host, select token types and locations, and maintain a record of deployed lures.
• The experiment connected canary tokens to a webhook-based email endpoint for alert ingestion.
• The presenter noted that the initial inventory appeared to cover only the current directory, rather than the entire host as requested.
• A generated QR-code token was placed on the virtual machine's desktop and successfully triggered the webhook when scanned.
• A management dashboard and Slack notifications were discussed as possible extensions, not demonstrated as completed features.