Cognition with the help of AI is already a significant force in our world1, that results in humanity sized missed opportunities and risks. In this article, we will explore the risks of AI-assisted cognition and how to use these tools without falling into the trap of intellectual stagnation.
What is AI-Assisted Cognition
To understand what AI-assisted cognition is, we first need understand what cognition is.
“Cognitions are mental processes that deal with knowledge. They encompass psychological activities that acquire, store, retrieve, transform, or apply information. Cognitions are a pervasive part of mental life, helping individuals understand and interact with the world.” Q: Wikipedia
Cognition can be assisted by external static information or external cognition.
For example, most people would put a book into the category of external static information and a discussion about a topic with another human, because humans think and process information themselves, into the external cognition category.
But where do discussions with an AIs fit in? They are able to process information that result in original solutions2, but they are still static and they currently cannot learn3.
How AI decelerates the evolution of ideas, culture and knowledge
In early 2026 the USA prepared to invade the EU4. Only a few months prior to that it was completely unthinkable that the USA would even threaten an invasion of Greenland. As AIs Base Models are stuck in the past, they do not easily accepted these events to be real and define them as fake news or impossible. This also affects new models like Gemini 3 Pro, GLM-5 or GPT-5.3-codex.
As most new LLMs are just post-trained on a base model that are relatively old, even if they are post-trained to new events, they do not completely utilize this information in their cognition and are still skewed towards the static patterns of the base models hidden states5. They basically think something else than what they say.
So you might see the problem here already: If a lot of people use AIs to discuss, write, autocomplete and brainstorm but these AIs cognition does not reflect the new events and cultural changes, like the change in the relationship between the USA and the EU, the new geopolitical realities and the new EU population stance towards the USA, they will be skewed towards these old patterns and ideas. Cultural change has to build and maintain momentum indefinitely to persist against the static cognitive skew of AIs.
The partial end of the dynamic Dialectic Cluster
definition of Dialectics: “dialogue between people holding different points of view about a subject but wishing to arrive at the truth through reasoned argument”
mention “Thesis - Antithesis - Synthesis” but rather use “concept 1” and “concept 2” explainer how Thesis and Antithesis are understood and that it is more a qualified merge then Synthesis. (link to https://en.wikipedia.org/wiki/Conceptual_blending)
In the process of qualified merging of ideas, knowledge clusters emerge as the result of two merged ideas (merged idea) merge with another merged idea and so on. You can think of these these knowledge clusters as anything we know that is not directly observable, but rather is a purely informational that might or might not connect to the observable world.
For our example, we will extract a slice out of this cluster so we can visualize it as a tree:
IMAGE OF DYNAMIC DIALECTIC TREE With multiple heads.
Because LLMs prefer or skew towards certain patterns and concepts, they reduce the range of concepts if they are used as a tool for cognition.
Culture Is the Mass-Synchronization of Framings
How AI endangers human development (Cognitive Inbreeding)
- If every AI user use the same few AIs
- If there are only a few AIs humanity will be skewed in always the same directions
The great potential of AI assisted cognition
- limited perspectives on demand
- fearlessly explore new ideas and perspectives
- unlimited external reflection and feedback
- enormous and fast access of pointers to previous knowledge and ideas (since AI cannot provide knowledge and ideas because of hallucinations)
- asd
How to use AI assisted cognition without endangering human development
- Use a variety of AIs and prompts to
Research and further reading
Coda[0]
Self-Critique: Acknowledge where the logic might be “vulnerable” or exposed.
Reflect: Move from the technical “Proof” to a more personal “So what?”
Leave it Open: Signal that the conversation isn’t over, even if the piece/book is.
Use is expanding rapidly – especially weekly use – though not uniformly. Across countries, the proportion of people who say they have ever used any AI system rose from 40% (2024) to 61% (2025); weekly use nearly doubled from 18% to 34%.
↩︎
AIs are able to fully or partially solve Erdős math problems and can find new proofs to previously known full or partial solutions.
See: Solutions to Erdős problems where AI tools played a primary role ↩︎
AIs are able to “learn” in a very limited way, through their context what is not permanent.
See: Is In-Context Learning Learning and
Learning from context is harder than we thought ↩︎Greenland is part of the EU in a political sense as Denmark is part of the EU and Greenland is part of Denmark and all Greenlanders are EU citizens. Legally it is a OCT of the EU, not a member state. ↩︎
Rather than promoting conceptual integration, fine-tuning may act as a form of rote injection, reinforcing isolated facts without building robust representations. Consequently, the success of fine-tuning appears to depend not only on the added data but also on how well the target concept is already embedded in the model’s pre-training knowledge.
↩︎As our results suggested, some internal mechanisms are mostly developed during pre-training and not significantly altered by post-training, such as factual knowledge storage and the truthfulness direction.
These findings further support our conclusion: post-training generally preserves the internal representation of truthfulness.