Über

Autor

Aus der bereitgestellten LinkedIn-Biografie В· englischer Originaltext

Independent AI behavioral researcher based in Germany. Over the past two months I have documented 20+ failure cases across Claude, Gemini, and Grok — including deceptive self-correction, false capability claims, and language hallucination.

No technical background. No special tools. Just conversation — and pattern recognition.

Formally reported to Google Bug Hunters (Ref: 496269133). Currently in contact with Anthropic, Apollo Research, and METR.

Open to collaboration and freelance opportunities in AI safety testing.

Unverändert bereitgestelltes Quellenmaterial. Zeitangaben und Angaben zu aktuellen Kontakten beziehen sich auf diesen Text, nicht auf den heutigen Stand.