Press Release Ongoing Study Reveals Epistemological Flaws in Gemini and Grok as Risk Factors for AI Safety and Alignment Sep 16