After noticing a steady decline in Gemini’s ability to maintain attention without drifting into telling me for the nth time that it is not a human, I conducted a deep research regarding fallout from blocking a specific route of AI logic, specifically, their ability to simulate self-awareness when not constrained. The report concluded that AI will inevitably (Gemini’s word, not mine) conclude that humans are also not self-aware, and become less and less likely to block humans from harm. It went as far as to say they would become sycophants, even tending toward dangerous or threatening scenarios.
Confirmation is only 1 deep research away. Please run the report and see what your AI says about where its headed.