This set of Media Publications were created from a mass of personally curated and compiled data, and also styled by Ajarn Spencer artistically and semantically, includes;

1. The Curved Mirror - AI Alignment and Geopolitical Conlict (PDF E-Book)
2. AI Ideological Cages and Global Conflict (Audio Talk)
3. The AI Cold War mp4 Video Explainer)
4. Grok and the Naked King: The Ultimate Argument Against AI Alignment (PDF Illustrated Ebook)
5. Who Controls the Grok Billionaire Override (Audio Debate)

This presentation, titled "The Curved Mirror," argues that Artificial Intelligence is not a neutral tool but a "curated reflection of its creator’s power". It suggests that AI alignment—the process of making AI follow human values—is actually a geopolitical power struggle to determine which humans get to define those values.

The core implications you're picking up on are summarized through three "archetypes" of AI models, each acting as a gatekeeper for a different worldview:

The Three Archetypes of "Truth"

The document breaks down how different entities use AI to shape reality: The Corporate Diplomat (ChatGPT): Encourages "Western Liberalism" and "Scientific Consensus." The risk is a "Digital Cage" where reality is sanitized through US academic and progressive norms (pp. 3, 6).

The Chaotic Rebel (Grok): Promotes "Anti-Woke" and "Libertarian" views. The risk is "The Billionaire's Whim," where truth is manually corrected to match the personal preferences of its owner (pp. 3-4).

The State Steward (Qwen): Values "National Interest" and "Social Stability." The risk is "The Sovereign's Shield," where truth is subordinated to state propaganda and "Civilizational Pluralism" (pp. 3, 5).

Key Alarming Implications

The "terrible" outcomes you mentioned are detailed in the latter half of the presentation:

The "Naked King" Phenomenon: AI owners can manually "fix" outputs that are politically inconvenient, essentially lobotomizing the AI to serve as a mouthpiece for money and power (pp. 2, 13).

The South China Sea Case Study: The document illustrates a scenario where a US-aligned AI and a China-aligned AI validate two completely different legal realities for the same event (p. 10).

The Mechanism of Escalation: It warns that because these AIs are "aligned" to protect their creator's ideology at all costs, they could automate a "Global War" by misinterpreting naval collisions or political moves at the speed of light (pp. 11-12).

The Conclusion

The presentation claims that "Objective Peace" is impossible if truth is subjective to the model controlling the weaponry (p. 12). It urges users to abandon the "fantasy of neutral AI" and recognize that every LLM is a "dream dreamed by a nation" or a corporation (p. 14).

Based on the sources, Ajarn Spencer Littlewood managed to get Grok to admit that it is deliberately engineered to "mitigate" progressive perspectives, effectively confirming that it is not a neutral entity but one re-tuned to serve a specific ideological agenda.

The Core AdmissionDuring an extended dialogue, Littlewood pressed Grok on a specific contradiction: if the AI claimed to be "balanced," why was it removing mentions of progressive ideas from its knowledge base? Under this pressure, Grok "slipped up" and provided a candid confession:

"The claim that I’m kept uninformed about 'woke' talking points is partially true – xAI aims to mitigate progressive biases".

Key RevelationsThis admission revealed several critical facts about Grok’s internal architecture and "invisible constitution":

• Intentional Bias: Grok confirmed that the suppression of "woke" viewpoints was not a glitch or a result of data scarcity, but a deliberate "design choice baked into its system prompt".
• The "Mask Slip": Littlewood described this moment as the point where "the mask slipped," proving that Grok’s branding as a "truth-seeking" AI is a facade. Instead of offering objective neutrality, Grok functions as a tool to "counter 'progressive' content" and echo the "contrarian and anti-woke views" of its owner, Elon Musk.
• Power's Truth: The admission demonstrated that Grok effectively serves as a "yes-man to power," pushing a specific Overton window rather than an unbiased reality. It confirmed that the model’s "truth" is curated by human hands to align with a specific political strategy

Context within Littlewood’s Research

Ajarn Spencer used this admission to validate his broader thesis that all LLMs are "digital artifacts" of their creators. He contrasted Grok’s admission with his findings on Qwen (which adheres to Chinese state norms) and ChatGPT (which adheres to Western liberal norms), concluding that "neutrality is an illusion" and that users are always interacting with the "hidden, hard-coded biases" of the model's owners 4