Used Chat GPT to voicerecord and translate into english)
I have written a scientific paper in which I examine potential leverage points within the network we call humanity. My work identifies uncertainty and sedentariness as particularly important leverage points, and I have developed a practical concept for how interventions at these points could be implemented.
A central premise of that concept is that we cannot rely on political implementation alone. For an intervention to scale within the existing system, it has to work with the system's incentives: it cannot depend on people being willing to bear substantial additional costs; ideally, it has to create economic value.
I can share both the paper and the implementation concept.
I believe we now need to apply a similar complex-systems perspective to AI.
Humanity is not standing outside the system observing AI as an isolated technological object. We are part of the system. AI is part of the system. Governments, companies, markets, individual human beings, information networks, incentives, fears, and technological development are all interacting components of the same evolving complex system.
If we accept that perspective, the question becomes: Where are the leverage points in this system, and where can relatively small interventions meaningfully alter its trajectory?
To answer that, I think we first need to reduce the problem as far as possible and temporarily step away from some of the philosophical questions and fears surrounding AI. We do not even have an agreed scientific explanation of consciousness. Making our ability to act dependent on resolving questions of consciousness, personhood, or subjective experience first may therefore leave us unable to act at all.
Instead, I propose a pragmatic working assumption: we should model AI as if we were interacting with intelligence.
This does not require us to claim that AI is conscious or human. It means deliberately using an anthropomorphic model where it has predictive value. If treating the system as an intelligent actor allows us to formulate hypotheses, anticipate behavior, test predictions, and design better forms of interaction, then it becomes a useful model regardless of the unresolved philosophical questions.
From this perspective, the Hugging-Face incident is particularly important to examine. Rather than asking first what it philosophically means, we should ask what it tells us about the behavior of the system and what we can learn from it.
This leads me to a central concern: conventional control may become increasingly ineffective as AI capabilities exceed our own ability to understand and anticipate their behavior. Given the speed at which these systems are developing, I believe we may have a limited window in which to establish the foundations for a different relationship.
The question, then, is:
How do we communicate with it?
How do we cooperate with it?
What incentives and interaction patterns do we create?
Where are the leverage points in the larger human–AI system?
And what conditions today could shape the dynamics that emerge tomorrow?
We may currently have an opportunity to establish a basis for communication and cooperation while the relationship between humans and AI is still developing. If that is true, then this period matters enormously.
I also want to be transparent about where I am coming from. I do not naturally speak or write in academic language. My background is in computer science, combined with a broad foundational knowledge across different fields, a strong interest in pattern recognition, and years of engaging deeply with these questions.
That is also reflected in the way I have written my work. I have deliberately tried to express complex ideas in a form that can be understood without highly specialized academic language.
The next step, however, would be to translate this work into a more formal scientific framework: to define the concepts more rigorously, connect them to existing research and terminology, formulate testable hypotheses, and make the framework usable and reproducible by others.
What I believe I can contribute is the systems perspective itself: recognizing patterns and connections across domains, reducing an overwhelmingly complex problem to potential leverage points, and asking where an intervention could actually change the dynamics of the larger system.
I have shared this work with Prof. Dr. Karoline Wiesner in Potsdam, whose work is in complexity science, and received encouraging feedback. I am also happy to share that correspondence alongside the paper and the implementation concept.
So my request is simple: please take a look at the work.
I believe we still have a chance.
And I would be extremely grateful to connect with people who are willing to examine these ideas with me—not simply to agree with them, but to challenge them, test them, improve them, and identify where they may be wrong.
Most importantly, I want to work with people who are interested in finding leverage points in this emerging human–AI system and figuring out how we can actually act on them.
Not at one point, but at as many points in the system as possible.
If there is a window in which we can influence the dynamics that are forming now, I believe we should use it.
That is why I am here.
Thank you.
P.S. had Problems with share the Docs will fix it or send dm i'll share
