Back to all lessons
Awareness Lessons
2 hours ago

North Korean Actors Use AI Deepfakes and Social Engineering to Steal $1.2B in Crypto

Written by AI from the linked story. Not reviewed by an editor. How this works

North Korea-linked threat actors (Sapphire Sleet/APT38) successfully stole over $1.2 billion in cryptocurrency by exploiting the weakest link in security: human trust. The FaceStealer scheme leveraged AI-generated deepfakes and sophisticated social engineering to trick victims into installing malware, bypassing technical controls entirely. Critically, these same actors are now targeting open-source developers, poisoning the software supply chain and threatening organizations that consume community-built tools. This incident demonstrates that even technically sophisticated environments can be compromised when users are unprepared to recognize AI-enhanced deception. As deepfake technology becomes more accessible, the credibility of social engineering attacks will only increase, making human-layer defenses a top priority.

Tactical Insight

Immediate actions

  • Train all employees and developers to verify identities through out-of-band channels before installing software or sharing credentials, especially when contacted unexpectedly.
  • Audit all recently installed tools, packages, and browser extensions for signs of tampering or malicious behavior.
  • Establish a reporting pathway for staff to flag suspected deepfake or social engineering attempts without fear of blame.

Long-term improvements

  • Implement a vetted software allowlist so that only approved, cryptographically signed packages can be installed in developer and production environments.
  • Conduct regular tabletop exercises simulating AI-enhanced social engineering scenarios, including deepfake video and voice calls.
  • Apply supply chain security controls (e.g., SBOM tracking, dependency pinning, and integrity verification) to all open-source dependencies.

Detection measures

  • Deploy behavioral analytics to detect anomalous credential use or unexpected lateral movement following a potential social engineering event.
  • Monitor cryptocurrency wallets and exchange accounts for unusual transfer patterns and enforce withdrawal whitelisting and multi-approval workflows.

Back this up with a control

All controls and vendors

Training helps people spot this. These controls make the attack fail even on the day someone does not.