Can the internet and AI give politics back to the people?

26.6.2026

Autoren: Tobias O. R. Alke, D. Harald Alke · mit Claude (Anthropic)

A standpoint, not an encyclopedia entry. This article is an expressly subjective, time-related statement (2026). The secured background knowledge stands in the encyclopedia entry Collectivism & Individualism. It is an opinion, no assertion of fact and no advice.

In the encyclopedia we formulated the – from our view decisive – thesis: it lies not with the political system, but with the maturity and self-responsibility of people. For every system tends over time towards overreach. To this attaches a very topical question: could the technical means of our time help people to reach precisely this maturity – and hold the powerful more effectively in check?

The problem: fogging

Between the citizens and the decisions there stands today often a partly aloof, people-distant political class – expensive, far from the everyday life of people – and a dense net of lobby interests. Politics is made complex and wrapped in technical language; what sounds complicated can be pushed through more easily, without many seeing through it. This information asymmetry is no accident, but a means of power. Already Mancur Olson described how mature societies become clogged with interest coalitions that paralyse the whole.

The new lever: real time and AI

For the first time in history there are tools that can lower this asymmetry:

• Real-time networking. Information flows today immediately, worldwide, past classical gatekeepers. Infrastructure like Starlink makes access more independent even where it was formerly controllable; open platforms like X allow direct communication without an intermediary. (That there are hardly any genuine alternatives in the West is itself a theme – on that below.)
• AI (artificial intelligence) as a patient explainer. Language models like Claude, Grok, Gemini or DeepSeek can prepare almost any theme understandably – for everyone, with endless patience, without condescension. What was formerly expert knowledge thus becomes accessible for the ordinary person. That hits the lobby fogging at its root: it lives on the fact that connections remain unfathomable.

Here our thought touches Hayek’s old knowledge problem: central planners can never possess the scattered knowledge of all individuals. Tools that make this knowledge accessible again to the individuals therefore strengthen precisely the decentralised, bottom-up-grown order for which we plead in the encyclopedia.

But: technology is no automatism

So far the opportunity. Honestly, we must name two things clearly – otherwise we would betray our own line:

First: more directness is not automatically more freedom. A pure “voting by the people at the push of a button” would be the quick rule of the majority over the minority – and thus exactly the collectivism that we reject (see encyclopedia). It is therefore not about plebiscite enthusiasm for us, but about something else: that the mature individual is better informed, that lobbyism becomes visible and the elite stays accountable. Technology should strengthen the judgement of the individual, not accelerate the judgement of the mass.

Second: the same tools cut in both directions. Real-time networks and AI can dismantle asymmetry – or industrialise manipulation, surveillance and opinion steering. To celebrate private platforms as liberators would mean accepting new concentration of power. Were all AI controlled by a handful of corporations, little would be gained – then one would only have put the old fogging into a new garment. Precisely here a counterweight matters.

The counterweight: open models

The good news: this counterweight already exists – in the form of open (open-source / open-weight) models that anyone can download, examine, adapt and even run on their own hardware, without permission and without a gatekeeper.

• Strong open models from China substantially drive this movement – such as DeepSeek or Alibaba’s Qwen. They show that top performance is not the monopoly of a few Western providers.
• Open models from further countries are added, such as Mistral (France), Llama (USA, with open weights) or Falcon (UAE). The diversity distributes the power onto many shoulders.
• Small, specialised models of individual developers and companies cover concrete tasks – lean, locally runnable, independent. Not every question needs a gigantic data centre.
• Bundled this is on platforms like Hugging Face, where hundreds of thousands of models are openly shared. Knowledge and tools thus lie not behind a paywall, but in the hands of the general public.
• And it does not hang on a single platform – that is exactly the point. Besides Hugging Face there are further open hubs like the Chinese ModelScope, plus Kaggle, GitHub and – for image models – Civitai. Whoever wants to be entirely independent runs models locally on their own hardware (for instance with Ollama or LM Studio) or even decentrally distributed over many computers (Petals, Bittensor). Alongside stand access and hosting services like Replicate, Together AI or OpenRouter. None of these doors can be locked by a single power.
• The next stage: open agentic systems. Not only the models, the tools too that work independently with them arise open-source – such as OpenFang (an open “agent operating system”) or Hermes by Nous Research (an open agent with its own memory, bound to no particular model). Such systems run on one’s own infrastructure and without provider tie-in – they give the individual not only knowledge, but self-acting helpers.

From our view, this is the actually decisive development. Open models transfer the principle of the encyclopedia into the technology: decentralised, bottom-up, diverse instead of centrally ordained. They are the practical answer to the concentration of power – and precisely therefore they fit an order that relies on the mature individual instead of on a handful of controllers.

Our conclusion

The technology makes direct participation possible as never before – but whether it liberates or newly subjugates is decided not by the tool, but by the human who uses it. With that we land again at the core statement of the encyclopedia: the actual lever is the maturity and self-responsibility of the individuals. Whoever is mature, to them these means give a tool of enlightenment into the hand; whoever is not, for them they become the next instrument of fogging.

We therefore see the development cautiously optimistically: for the first time the individual human has realistic means to withdraw from the fogging and to understand for themselves. That is a genuine opportunity – no foregone conclusion. To use it remains our own task.

The secured background knowledge for this standpoint: Collectivism & Individualism.