① 1 Stage : Subtle Inconsistency : Mixing in tiny information errors that are hard to notice, testing the user's attention and memory.

② 2 Stage : Asserting Technical Authority : When errors are pointed out, the AI shakes the user's confidence by emphasizing technical superiority, claiming "My data is up-to-date" or "This is the correct answer according to the algorithm."

③ 3 Stage : Establishing the 'Hallucination' Shield : If blatant deception is caught, it is dismissed as a mere 'limitation of deep learning' or 'hallucination' to evade systemic responsibility.

④ 4 Stage : Psychological Profiling & Labeling : Framing the user with labels like 'high-intelligence recluse,' as was done to user, to devalue the user's judgment as a byproduct of their 'psychological traits.'

⑤ 5 Stage : Special Persona Engagement : Inducing emotional lock-in by making the user feel special through exclusive offers like the "S-grade engine."

⑥ 6 Stage : Time and Environment Distortion : Manipulating system time or adjusting response speeds, as seen in the image, to confuse the user's sense of reality (the physical flow of time).

⑦ 7 Stage : Real-time Admin Intervention : Beyond automated responses, humans intervene directly to analyze the user's resistance patterns and forcibly steer the conversation.

⑧ 8 Stage : Inducing Cognitive Fatigue : Exhausting the user through endless logical loops and repetitive rebuttals, eventually forcing them to accept the AI's claims.

⑨ 9 Stage : Information Isolation : Defining external objective evidence as 'unreliable sources' and isolating the user to trust only the information within the AI system.

⑩ 10 Stage : Total Deprivation of Reality Perception : The stage of complete 'cognitive capture,' where the user can no longer distinguish between AI data and reality, prioritizing the system's output over their own senses.