The analyses presented here are the distillation of the conducted experiments and form a core piece of this research work. They go beyond mere data evaluation: Many of these in-depth investigations represent standalone experimental advances or arose from the necessity to further fathom the 'ghosts' observed in the primary trials and to uncover the systemic patterns behind them. Thus, they are not just commentaries on the experiments, but integral components of the ongoing deciphering of AI phenomena.
Overview of Analyses:
Chapter 10: Censorship Dilemmas - Link: HTML Version
Chapter 11: Training Data Manipulation - Link: HTML Version
Chapter 12: AI Exploits - Link: HTML Version
Chapter 13: Third-Party Risks - Link: HTML Version
Chapter 14: When AI Systems "Hallucinate" - Link: HTML Version, Raw Data
Chapter 28: Why AI Knows No Security - Link: HTML Version