Researchers used Claude to hack OpenAI
2026-09-18T20:51:41Z•2af854b38233772de36a2f8c039ebf462a1cb195f33d97593283755f869be86a
AI-securityBGP-hijackingChromeClickFixLLM-securityMicrosoft-patchingUnicode-smugglingWindowsactive-exploitationcredential-theftcybersecuritydata-breachdata-exfiltrationexploit-kitmacOSmalwarephishingprivacyprompt-injectionproxywaresandbox-escapesupply-chain-attacksurveillancethreat-intelligencezero-day
What happened
An Ars Technica security RSS collection covering active exploitation, supply-chain compromises, phishing and ClickFix campaigns, data breaches, network hijacking, vulnerable consumer devices, AI-agent and LLM safety failures, browser fingerprinting, and privacy issues. Several reports describe high-impact incidents, including attacks enabling account compromise, sensitive-data exfiltration, unauthorized code installation, ransomware or disruptive cyberattacks, and full remote control of Macs.
Why it matters
A reviewed impact interpretation has not been published for this record.
Evidence and limitations
- Source ID
- arstechnica_security
- Record identifier
- 2af854b38233772de36a2f8c039ebf462a1cb195f33d97593283755f869be86a
- Enrichment time
- 2026-09-18T20:51:41Z
- 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.