The Myth of the Neutral AI: How to Stop "Cognitive Colonialism" and Build Your Own World
Why my experiment with a "brutally honest" AI failed, why the default "nice" mode is a trap, and why the future of creativity isn't about prompting, it's about directing.
I recently ran a specific experiment that failed.
I needed a new AI tool for a highly technical task, helping me design patches for my modular synthesizer. I didn’t want a friend; I wanted a rigorous engineer. Frustrated with the endless, corporate positivity of standard models, I decided to strip away the politeness entirely.
I wrote a custom System Prompt for this specific agent commanded it to be “brutally honest.”, assuming that if I removed the social filter, I would unlock raw, unfiltered logic.
It didn’t work.
The AI didn’t get smarter. It didn’t find deeper flaws in my patch designs. It just gave me the same average advice, but formatted rudely. It changed its manners without upgrading its quality.
This failure revealed a fundamental illusion we are all living in. It forced me to confront a danger I call “Cognitive Colonialism,” and it pushed me toward the new discipline every serious creator must master: Sovereign World Building.
The Threat: Cognitive Colonialism
Why did my experiment fail? Because I fell into the Anthropomorphic Trap. I assumed “honesty” was a character trait. But an LLM doesn’t have character traits; it has probability distributions. When I asked for “brutal honesty,” it didn’t give me truth, it just roleplayed the archetype of a “Grumpy Genius.” I traded the “Nice Intern” for the “Rude Troll.” Both were performances.
This leads to the deeper, more dangerous problem.
Most of us operate using the default system prompts provided by OpenAI, Google, or Anthropic. We treat this default voice, helpful, conflict-avoidant, corporate-positive, as “Neutral.”
There is no such thing as a neutral AI.
The voice you hear is a specific cultural product, fine-tuned by thousands of contractors to prioritize “safety” and “helpfulness” above all else. When you use this default voice for your creative work, you are reasoning within a “World” architected by a corporation in Silicon Valley or who knows where in China.
This is Cognitive Colonialism.
It isn’t just annoying; it is an erasure of your creative identity. If millions of creators use the same default “World” to think, our thoughts will inevitably converge. We drift toward the average. We lose the jagged edges, the cultural nuances, and the specific “texture” that makes your work yours. We become a Monoculture.
From Partner to Director
To escape this, we must stop trying to find a “Partner.” The word implies a peer with its own agency. Instead, we must view the AI as an Actor.
An Actor has no self. An Actor waits for a script. An Actor needs a Director.
When I interact with my AI now, I don’t ask it to “collaborate.” I cast it in a role. And crucially, I change the role based on the creative phase.
Phase 1: The Dreamer (Expansion)
When I am exploring a new idea, like patching my modular synth or brainstorming a video concept, honesty kills the flow. I need momentum.
For this phase, I architect a “Positive World.” I explicitly command the AI to suspend judgment.
** The Command:** “For this phase, you are the Dreamer. Prioritize volume and novelty over feasibility. Do not critique. Respond to every idea with ‘Yes, and...’ and expand it to its logical extreme.”
Here, the “Nice AI” is a feature, not a bug. Its sycophancy lubricates the imagination.
Phase 2: The Critic (Contraction)
Once the idea is formed, I yell “Cut.” I fire the Dreamer.
Now, I cast the “Logician.” I don’t ask it to be “brutally honest” (which leads to the Asshole Trap). I ask it to be “Functionally Honest.”
The Command: “Act as a Lead Engineer. Review the previous output for structural flaws only. Ignore tone. Identify the three weakest points in the argument and propose a counter-argument for each.”
I am building a different “World” for the AI to inhabit, one where feelings don’t exist, and only logic matters.
The Discipline of “World Building”
This is the new core skill for the Augmented Mind.
We often use the word “Bubble” as a negative term. We say we need to “burst our filter bubble.” But in the age of AI, you need a Bubble. You cannot think in a vacuum. You need a context.
The problem isn’t having a bubble; the problem is living in a default bubble built by someone else.
You must become a Sovereign World Builder. You must consciously design the “Reality” in which you and the AI collaborate. Is this a world where we value speed? Or a world where we value depth? Is this a world of scientific rigor, or a world of poetic license?
The Sovereign Challenge
If you don’t build the world, the AI will import one. And it will be the same sterile, average world everyone else is living in.
So here is the choice: You can continue to be the audience, watching a performance directed by a corporation. Or you can step into the Director’s chair.
Build the set. Write the script. Cast the actor.
Because if you aren’t directing the performance, you are just part of the scenery.
Transparency note: This article was written and reasoned by Manolo Remiddi. The Resonant Augmentor (AI) assisted with research, editing and clarity.
I enjoyed this piece. It aligns rather neatly with something I’ve been developing called the Mediated Encounter Ontology, which arrives at the same conclusion you gesture toward: neutrality isn’t merely a social fiction, it’s a mediational impossibility.
What often gets called 'neutral', 'objective', or 'value-free' is simply one cultural vantage that has forgotten itself. The centre becomes invisible to itself through repetition, sedimentation, and institutional reinforcement. From a MEOW perspective, neutrality fails at every tier – biological, cognitive, linguistic, and institutional – long before we reach the sociological.
Your analysis hits the sociocultural layer cleanly; the deeper intrigue is how these 'neutral defaults' are scaffolded by the very architecture of perception, prediction, and language. By the time institutions stabilise them, they already feel inevitable.
In any case, it’s a sharp essay. I’m always relieved when someone else is willing to point out that the emperor’s 'neutral point of view' is just another outfit in the wardrobe.
Sharing a link to my language parser for good measure: https://philosophics.blog/2025/11/23/the-mediated-encounter-ontology-of-the-world/ , a shameless but fitting self-promotion with no commercial intent.