AI Tools Are Helping Mediocre North Korean Hackers Steal Millions
2026-04-23T08:52:13Z•3dd6458f8c611f08ea51c969fd7f266e5730f3ce70c8a23603d271ffedd3cf86
AI-assisted hackingAnthropic MythosDDoS attackMozilla/Firefox bugsNorth KoreaOpenAI GPT-5.4-CyberSection 702Telegram illicit marketplacesage-verification appcrypto scamsdeepfakes/sexual abusefalse advertising / scam ads (Meta)incident responseprivacy tools (GrapheneOS)supply-chain/software-securitysurveillance abusesweb application vulnerability
What happened
Collection of WIRED security stories (Apr 8–22, 2026) highlighting rising AI-enabled cyber threats, widespread software vulnerabilities, and large-scale privacy/abuse risks. Key items: North Korean-affiliated hackers using AI to automate malware, social engineering, and fake sites to steal up to ~$12M; Anthropic’s Mythos and other LLMs are being used both offensively and defensively (Mozilla used Mythos to surface and fix 271 Firefox bugs; OpenAI released a cybersecurity-focused GPT-5.4 model); an EU age‑verification app was trivial to hack; a global surge in AI deepfake nudes affecting school
Why it matters
A reviewed impact interpretation has not been published for this record.
Evidence and limitations
- Source ID
- wired_security
- Record identifier
- 3dd6458f8c611f08ea51c969fd7f266e5730f3ce70c8a23603d271ffedd3cf86
- Enrichment time
- 2026-04-23T08:52:13Z
- AI-assisted enrichment
- Yes
This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.