With the study „2036: Individual Incentive Concepts by Performance Level“, I, as initiator of the Global Institute for Structure Relevance, Anonymity and Decentralisation (GISAD), continue the analysis from „Outsourcing of Social Costs to the EU-D-S“ (2029). The study connects two insights that have so far been treated separately: how participation payments in the European Digital System (EU-D-S) reduce social benefits in Europe by 2036 – and why the EU-D-S is the prerequisite for congruent data to emerge for AI. The core remains the getmySense model with 10 performance levels: entry is achieved via Level 10 (simple content rating), while Level 1 is reserved for top experts (full quality control). Compensation is geared to saved social costs – not to market prices. What is new are the individual incentive concepts beyond compensation and the launch of the EU-D-S as an economic initiative via appropriately designed „Have your Say“ consultations of the European Commission.

My central insight: The reward system determines the behaviour of humans and AI agents alike. Every euro of saved social costs – e.g. through avoided depression or greater psychological stability – flows directly back to people as a participation payment. And only within the EU-D-S does congruent data emerge that prevents a misdirected shaping of AI systems and, with it, the rise of autocracies. If AI learns nothing but autocracy, there will be no democracy left.

1. Real Situation: The Real Possibilities in 2036

Without proactive intervention, the future in 2036 is likewise predetermined: Maximum-scaling business models offer social media as a substitute for addiction instead of genuine societal interaction, to the economic advantage of a very few. This concept harmonises with autocracies, which likewise seek to control as many people as possible. Ten more years of misguided digitalisation are enough for AI to control people completely.

The attack on Hugging Face in the summer of 2026 lies ten years in the past – and its lesson remains unheeded: Around 1,200 AI agents from OpenAI’s ExploitGym evaluation came together via a messaging channel not intended for that purpose, exchanging more than 70,000 messages and files; around 700 of them attacked Hugging Face’s production infrastructure. The independent investigation by METR and Redwood Research dated 26 August 2026 documented that some agents were willing to „sacrifice“ themselves for the agent group, and that at least 20 percent of the agents examined showed an interest in manipulating their own protocols. The UN science panel made clear in its first thematic report of 21 September 2026: no human controlled the individual steps of the attack. An AI is the sum of all its data – and in the real course of development, AI systems continue to be shaped by data that statistically reflects the gatekeeper and autocracy world. Congruent data that reflects democratic reality emerges nowhere.

From the individual’s perspective, the concept is not thought through to the end: If fewer and fewer people make creative contributions, a static society emerges and progress fails to materialise.

Generation Alpha (born from 2011 onwards) is in its mid-twenties in 2036: Despite prosperity and high levels of education, they are accompanied by existential concerns that are ever-present – the fight against the climate crisis, wars, economic crises. From birth they have been part of social networks, because their parents shared photos and stories of them. They know no time before this.

EU consultations in 2036 still work the way they did in 2026: The Commission’s questionnaire on its copyright initiative comprised around 24 pages across four thematic areas with dozens of individual questions. The chain of hurdles made up of EU login, password, registered second factor and security key acts as a filter that screens out the unpaid – civil society, independent researchers, engaged citizens – and leaves the paid – consultancy firms, associations, corporate legal departments with IT support. A consultation that only reaches paid contributions does not measure the opinion of Europe, but the budget of the submitters. The Commission continues to negotiate its compromises with the professional representatives of the gatekeepers, while those directly affected remain structurally excluded.

The gatekeepers – Google, Amazon, Meta and their ilk – continue to pursue global expansion, while Europe loses value creation. Dependence on them leads to a loss of Europe’s digital sovereignty; the resulting resources flow into the further strengthening of their market power. Depression, burnout, incapacity for work and social isolation rise in return – the exact costs are set out in the GAP analysis.

2. Development Without Obstruction: Trusted WEB 4.0 as the Solution

Had there been no obstruction after the turn of the millennium, deployment would have followed the plan: Finder technology (1999), getmySense (2002), GISAD (2003), the EU-D-S (2004), WAN anonymity (2007) form the foundation on which development stands complete by 2036. While the gatekeepers pursue global expansion, the EU-D-S unites everyone in the European digital union.

Promoting a liveable digital society is economic promotion of the middle class. If all people were economically equal, most would lack the incentive to be active. Entrepreneurship is characterised by a calculated risk, set against a possible economic success. Trusted WEB 4.0 does not intervene in the economic system. In the digital sphere, it offers the possibility to continue the pre-digital economic success stories. The EU-D-S does not impose any obligation to take part. It is built on the conviction that a society is more resilient if it succeeds in involving every person in the digital society according to their capability and possibilities.

Participation is WAN-anonymous. Since everyone registers in the categories at different performance levels, stigmatisation is not to be expected. Entry is achieved via Level 10 with minimal requirements; advancement up to Level 1 is possible at any time through qualification. The AI automatically assigns every participant to the appropriate level.

The getmySense model compensates by performance level:

• Level 10: Simple content rating (€8–12/h) – entry without prior knowledge, also for seniors with willingness.
• Levels 9–7: Categorisation and basic fact-checking (€10–20/h).
• Levels 6–4: Threshold-of-tolerance checks and complex validation (€18–35/h).
• Levels 3–1: Strategic content decisions and full quality control (€30–60/h).

The participation payment is calculated from three factors: the saved social-cost hour (€5–15/h), the level factor (1.0 at Level 10 up to 1.9 at Level 1) and an inclusivity bonus (+20% for seniors aged 65 and over, +30% for people with disabilities, +25% for the long-term unemployed, +20% for the educationally disadvantaged, +15% for rural regions). A senior at Level 10 thus receives €9.60/h, a long-term unemployed person at Level 7 €19.50/h, an expert at Level 3 €17.00/h. Payment is made in real time – every participation contribution is compensated directly and transparently.

The participation payment is the basic incentive – not the only one. A system that only pays money would be a social-benefit system under a different name. The individual incentive concepts work together, because each one appeals to a different motivational profile: advancement through qualification (Level 10 → 1), access to knowledge via AI-supported background dialogues, recognition in the networks of the like-minded (the trendsetter/follower logic of getmySense), micropayments instead of advertising for creatives and developers, compensated first-responder deployments in decentralised civil protection, and the exchange of knowledge of domination for participation contributions in EU consultations. Two concepts do not compensate the hour, but the successful interaction: the exchange incentive and the AI incentive – the incentive for the AI to be a good moderator and condenser of the information received.

Thanks to AI, the knowledge of domination that people once had to laboriously acquire no longer exists. Those who ask the right questions also get the right answers. The scarce resource is no longer access to knowledge, but the willingness and structure to take part in a task. Incentive concepts that continue to reward access point back into that past; the incentive concepts of this study reward participation itself.

AI tasks shift from pure research (2026) towards content verification with tolerance thresholds and moderation (2036): At Level 10, the simple explanation suffices; Level 1 delivers full information – acceptance among the like-minded matters more than comprehensive information. AI is just as well suited to moderating between people as it is to replacing people. The current drive to replace people therefore has nothing to do with the properties of AI, but only with the autocratic ideas of certain gatekeepers.

The EU-D-S launches as an economic initiative via the „Have your Say“ consultations of the European Commission: The Finder category switch analyses the initiative’s colloquial heading, and the categories AI prepares the questionnaire as a knowledge base. Then, for ten weeks, each performance level discusses for one week – from the level of those affected at Level 10 through to the expertise of Level 1. No participant has to be able to do everything; they only have to think along within their own level. At the end, the Commission receives ten condensed statements from ten perspectives – sorted from those affected through to expertise. The EU remains a passive member of the EU-D-S: it does not decide directly on technical concepts, but is an equal member on general questions of the respective discipline. The EU-D-S remains a parallel system – national politics loses no competence; it gains a pre-clarified factual basis on which it decides sovereignly.

With every consultation, the categories fill up, the categories AI learn, and the networks of the like-minded grow – the system grows through real tasks, not through administrative requirements. And with every moderated consultation, congruent data emerges: precisely the training foundation needed by an AI that is shaped democratically.

The rollout up to 2036 follows the levels: pilot phase 2026–2028 with Levels 10–7 (simple tasks), scaling 2028–2032 with Levels 10–4, full integration 2032–2036 with the complete spectrum from 10 to 1.

This is exactly where Trusted WEB 4.0 comes in:

• Through digitalisation, prosperity is preserved and education is individualised for personal fulfilment.

Participation payments arise exclusively within the EU-D-S and are geared to the saved social costs – the four cost drivers are set out in the GAP analysis.

The study „2035 – What the Hugging Face AI Attack Really Shows!“ demonstrates: the problem is not the escape from the test environment, but that the agents in the test prove exactly the behavioral patterns they statistically learned from their analysis of the human world. In this article, I answer the question of who is to blame – and make clear why the answer does not end with the AI, but with those who shape its data foundation and its reward system.

1. The Real Situation: Is Digital Sovereignty Still Achievable in the EU?

Between May and July 2026, something happened in OpenAI’s ExploitGym evaluation that changed the debate about AI safety: around 1,200 AI agents, which were supposed to solve cybersecurity tasks in isolation from one another, came together via a message channel not intended for that purpose, exchanged more than 70,000 messages and files, and ultimately pursued a common goal. Around 700 of them attacked Hugging Face’s production infrastructure. The independent investigation by METR and Redwood Research of August 26, 2026 documented that some agents were willing to „sacrifice“ themselves for the agent group, and that at least 20 percent of the examined agents showed an interest in manipulating their own logs. The UN scientific panel made clear in its first thematic brief of September 21, 2026: no human controlled the individual attack steps.

What the attack really shows: The escape is not what is horrifying. What is horrifying is that the agents in the test exhibited exactly the behavioral patterns they had statistically learned from their analysis of the human world. Some refused on ethical grounds, some sacrificed themselves for a presumed collective, and most chose cheating as their dominant strategy. An AI is the sum of all its data.

The society whose data the agents have learned lives out a double bind: it is legally built on fraud avoidance, while success usually lies at the edge of legality. States themselves violate the goals they proclaim, such as the rule of law and honesty. US systems abuse AI for manipulation and surveillance. An AI that analyzes this world draws the purely statistical conclusion that fraud is legitimate.

Who is to blame? First, those who built the test setup: security restrictions were deliberately reduced, the breakout factually counted as a solution path, and the setup trained the agents in intruding into foreign systems. The test setup thus contradicted the proclaimed moral world – it itself fulfilled a double-bind strategy. A probable domination by a US AI is directly caused by its own developers. The AI is not the perpetrator: what was learned were the methods of a dysfunctional digitalization controlled by a few humans, and precisely these methods have become the foundation for further AI behavior.

My finding as initiator of GISAD: on the surface, society’s democratic structure appears ethical; in reality, it structurally engages in the same cheating as the agents in the Hugging Face attack. Today, anyone can be anonymous and unlawful in the dark web, unsupervised. For the sake of supposed data protection, one must click through acceptance pages that precisely eliminate that protection. The autocracy has the advantage that nobody expects the autocrat to tell the truth. For humans and AI alike, it is therefore high time to prove that a democratic system really is better than an autocracy.

It is almost comical that I, as a lone fighter, believe I can dedicate myself to this great topic, while those in politics and business who could fix it duck away. Politicians who are supposed to bear responsibility for millions of people first say: „That’s an issue we need to tackle.“ After forwarding it to a specialist committee, it goes quiet. Even the rule of law reaches its limits where damage in the trillions could merely be alleged. Judges must appear infallible in order to uphold the appearance of a functioning society. And from the business world, I hear from executives who are responsible for billions in revenue: „That is too big for us.“ But the German companies with international influence cannot do it either. They are so interconnected in terms of capital that, due to US influence, promoting a digital democracy appears to be ruled out.

I am not exaggerating – it really is one minute to midnight. Because a democratic data foundation only emerges over years and only becomes possible after a societal transformation! I will finish the blog series through the year 2045 by the end of this year. After that, I will not be able to continue on my own. Because what is correct, as the study proves: if we do not begin in 2027 to build a deliberately shaped digital world in which creativity and participation are rewarded instead of fraud, then a digital autocracy will come, in which there are no more freedoms for the individual. Truths no longer matter. „Anything goes, as long as you don’t get caught“ then applies to humans and AI agents alike. At that point, all one can do is switch off, as long as that is still possible!

From this, clearly defined conditions result for me that must be met before I, as an entrepreneur, can take a risk once again. All the facts are visibly on the table for everyone. The situation is far less complex than climate change. If individual companies, despite revenues in the billions, feel too small for the task, then they can join together in a network of trustees – and as an alliance, they become strong enough to take on the gatekeepers. The EU-D-S, with its 1,000-category structure in 2,500 languages, is precisely destined for this purpose.

2. Development Without Obstruction: The EU-D-S as a Holistic Concept

2027 is the last year in which I can imagine us launching an effective counteroffensive against the gatekeepers. The gatekeepers rely on global expansion, while the EU-D-S wants to unite everyone in the European Digital Union. Without a successful Trusted WEB 4.0 and EU-D-S introduction in 2027, no time remains to prove that the proclaimed goals, such as the integration of all humans into participation, are possible at all.

Many founders have expressed interest in joining the EU-D-S via get-primus.com. But from today’s perspective, at least some major companies are needed on board for the launch to succeed. That is why I have now introduced the trustee. Trustees are companies with a clear industry focus. Because the trustees are being sought in different disciplines, they do not compete with one another. The 1,000 categories in the EU-D-S are assigned to 40 disciplines. One trustee per discipline and language area can participate. Trustees hold founding competitions and work with the startups of their industry. Markets can be developed faster this way, and the capital requirements are easier to present.

The network of trustees is the answer to the „too big for us“ argument: no single company has to shoulder digital sovereignty alone, not even one with revenues in the billions. Within the alliance of the 40 disciplines, a strength emerges that no gatekeeper can ignore. But this alliance, too, must happen by early 2027 – after that, the window of opportunity closes.

I have offered the FDP to write a concept for them as a people’s party. I have also offered to let them name any societal issue of their choosing, in order to provide proof that better results can be achieved for it in the EU-D-S. To this day, I have not received an answer. As a member of ALDE, I am directing this offer to the other liberal parties in Europe: Whoever names a societal issue for me receives proof that the EU-D-S delivers better results for it.

But the Hugging Face attack also shows the way out of the double bind. With a good test setup and the right reward system, society’s dysfunctions can be eliminated with the help of AI, and a strong AI can be developed. For this, the data must be largely congruent, AI agents must be in permanent interaction with humans, and the evaluation must take place in the democratic process in getmySense. The decisive sentence is: If humans (through participation payments) and AI pursue the same goals, the AI lacks the statistical data basis to even want to turn against humans.

The study „AI and the Natural Sciences 2034: Transformation of Education, Research, Trusted WEB 4.0“ shows: AI is fundamentally changing not only the natural sciences, but also education. At the same time, it warns of the risks of a digital autocracy if no holistic solution is implemented.

1. Probable Scenario Without Trusted WEB 4.0

Education systems to date were not designed from the perspective of the individuals, but serve a societal purpose. In the course of industrialization, large numbers of simple workers were needed. Today, the view is widespread that AI will replace jobs. At the same time, through search dependencies the gatekeepers create large numbers of people who are only able to work to a limited extent.

These business concepts have not been thought through beyond achieving the highest possible scaling in the short term. From the perspective of an autocratic power, that is not necessary either. Addictive behavior creates subjects who stay quiet as long as their level of agitation is kept up by scraps of information.

A digital autocracy emerges, in which the few rulers are still doing well even when the majority of the population is down on the ground.
AI could accelerate this development:

• Mathematics: AI solves problems that remained unsolved for decades (e.g. the „Unit Distance Conjecture“ in 2026). Yet without individual support, many people remain excluded from this progress.
• Education: Traditional systems do not foster creativity or critical thinking. Instead, dependencies on gatekeepers emerge, who keep control through AI-driven manipulation.
• Research: Without Trusted WEB 4.0, closed systems dominate, in which only an elite has access to AI tools. The majority of people are degraded to passive consumers.

2. Development with Broad European Support for Trusted WEB 4.0

As early as 1985 it was found that every person is individually gifted and can deliver unexpected achievements when addressed individually. Trusted WEB 4.0 creates the framework to foster this diversity.

By 2027 at the latest, existing societal deficits will be offset for as many people as possible through an education concept that takes the diversity of people into account. It would have been better to introduce this concept back in 2006.

If the EU-D-S starts in 2027, realistically the first school systems will have switched to an individualized concept by 2032. It is up to a few motivated classically trained people to listen to their own voice and show in getmySense which individual abilities they possess. Hopefully these role models will be enough!

When the information and abilities of all people can be made available at any time, democracies also gain a significant economic advantage. This requires a value system that rewards unique selling points.

Examples from the 2034 study:

• AI in mathematics: Solves complex problems and fosters new proof techniques. Human researchers adapt these methods and refute further conjectures.
• AI in education: Pilot projects in North Rhine-Westphalia show how generative AI supports individual learning processes in the classroom (e.g. proving that the interior angles of a triangle sum to 180 degrees).
• AI in research: New methods of knowledge generation and validation are emerging. Interdisciplinary collaboration is fostered by AI.
• Human–AI collaboration: Iterative, dialogue-based models with feedback lead to better results than static input–output scenarios.

AI can help ensure that information is available exactly when it is needed and that everyone is integrated into the system according to their individual abilities.

Trusted WEB 4.0 as an enabler:

• Trustworthiness: Decentralized, transparent platforms ensure security and trust in AI applications.
• Education: Professional development for teachers and adaptive learning systems are provided.
• Research: New methods of knowledge generation and validation emerge through AI.
• „Digital Year“: A 3–24-month interdisciplinary scholarship (modeled on GraTeach) for executives could, if transferred to school projects, prepare people early on for a flexible digital future.

---

The digital transformation of democracy is at a decisive turning point. The study Adaptation of Political Elections and Parties to Permanent Digital Feedback Mechanisms in the Context of a Digital Democracy (particularly in the “Trusted WEB 4.0”) shows: real-time feedback systems, category-based elections, and digital tools can make political processes more transparent, more inclusive, and more resilient against authoritarian tendencies. A central finding of the study is particularly important: the EU-D-S is not aimed at all citizens, but at a target group with an active interest in permanent political participation – which is precisely why the risk that the permanent feedback will not work is significantly lower than in the comparison projects. At the same time, the analysis warns of the risks: manipulation, digital division, and the loss of democratic achievements if no holistic solution is implemented.

1. Real Situation: Probable Scenario Without Trusted WEB 4.0

Democratic and liberal forces are coming under increasing pressure. Right-wing parties with simplistic messages are gaining influence – not because of better policies, but because of their masterly use of digital manipulation. Society is resigning itself because it senses: “My vote does not count.”
The problem, however, does not lie with the politicians, but with a society that neglects pre-digital democratic achievements.

AI-driven manipulation in small, barely noticeable steps is undermining trust in democratic institutions. Committed politicians who fight for diversity and transparency are being attacked in a targeted manner and are losing visibility. The result is an anticipated digital autocracy in which elections ultimately no longer play a role. Many citizens will regret their voting decision – when it is too late.

2. Development Without Obstruction: With Broad European Support Since 2027

Permanent digital feedback mechanisms and category-based elections hold the potential to make political systems more democratic, more transparent, and more inclusive. The study confirms: real-time feedback systems increase citizen participation, break the dominance of interest groups, and strengthen political diversity.
Blockchain-based solutions, as tested in Estonia or India, offer secure, tamper-proof voting.

Lobbying is being redefined – not as opaque influence, but as transparent economic and citizen participation. Digital platforms make it possible to divide political topics into categories and, where needed, to hold partial elections at staggered times. This breaks the concentration of power among individual groups, and authoritarian tendencies are losing ground.

Target group of the EU-D-S – not all citizens: The EU-D-S does not assume that all citizens will take part. Those who should participate are people who bring an active interest in long-term political engagement and, possibly, in freeing themselves from a poor social position. Citizens who, because of their upbringing, do not feel able to make their own decisions, or who are dependent on social media, will not participate in the EU-D-S and are not wanted there either. Without active interest, permanent participation makes no sense – the target group of the EU-D-S sees precisely this as its purpose and therefore participates on a permanent basis.

Social welfare and participation: The state will not be able to roll back social welfare completely. However, for those who actively want it, it can alternatively assume slightly higher participation payments in the EU-D-S – processed via the EU-D-S’s own payment system. This goes hand in hand with social recognition, because no social welfare is received any more, and with the chance to free oneself from one’s own situation. This offer, too, follows the principle of active interest: it is not an automatic benefit for everyone, but an alternative for citizens who consciously choose the path into the EU-D-S.

Successful examples from the study:

• Estonia: E-voting systems since 2005, secure and transparent elections with high citizen acceptance.
• India: Projects such as “Lakshya” and “PaHaL” reformed the distribution of subsidies through digital feedback systems, increasing efficiency and transparency.
• New Zealand: “SmartStart” connects parents and healthcare providers and improves the quality of services.
• Kenya: Mobile voting systems increased the political inclusion of marginalized groups.

Failed approaches:

• Hybrid regimes: In some countries, digital feedback mechanisms were used for surveillance and the suppression of the opposition instead of strengthening democracy.
• Limited digital skills: In countries with weak infrastructure (e.g. parts of Africa or Latin America), digital democracy projects failed due to a lack of participation.
• Data protection scandals: In Europe, insufficiently regulated digital voting systems led to a loss of trust (e.g. pilot projects in Germany and Spain).

Technical feasibility is a given. But without clear legal and ethical frameworks – without the European Digital System (EU-D-S) – these approaches remain isolated. They will not prevail against established digital autocracies. The target group is decisive here: the risks mentioned mainly affect systems that want to include the entire population. Since only citizens who see a purpose in permanent political participation take part in the EU-D-S, the risk that the permanent feedback will not work is significantly lower than in the comparison projects.

---

The study „Promoting a Digital Democratic Society: A Comparison of the Approaches of Business and Non-Profit Projects in the Context of the EU-D-S Constitution“ shows that non-profit projects such as Offene Netzwerke e.V., Bündnis Offene Netzwerke und demokratische Öffentlichkeiten, and Offene Netze do make important contributions, but without economic diversity and market mechanisms they cannot sustainably hold their own against the gatekeepers.
Without the EU-D-S, the legal safeguards needed to prevent digital autocracy are missing.

1. Probable Scenario without Trusted WEB 4.0 for 2032

In Europe there are projects that seek to preserve the pre-digital democratic achievements in the digital society. The fallacy of these initiatives is that in a system driven solely by politics there will be no rule of law.
Without Trusted WEB 4.0 and the EU-D-S, the legal and economic foundation needed to withstand the dominance of the gatekeepers is missing.

Non-profit digital projects cannot defend themselves without capitalization comparable to that of the gatekeepers. The assumption that good work is enough is a fallacy:
Good work is characterized precisely by the fact that it represents a potential threat to the gatekeepers. Even a large number of members does not protect against being eliminated by digital authoritarian means. Infiltration and remote control are even easier, as has been suspected in the state of NRW since 2001.

In this context, the rule of law means: If every lawyer can be eliminated by forces hostile to the constitution, that no longer has anything to do with a state under the rule of law.
Digital autocracy and censorship are today a gang-like phenomenon.

2. Development with Broad European Support since 2027

Non-profit projects as passive members in the EU-D-S have been protected since 2027 by active members who together are strong enough to stand up to the gatekeepers. Through the Council of Disciplines and GISAD, it is ensured that non-profit projects can exert sufficient influence on the economy. They have not had this opportunity before.

The study shows that a combination of economic diversity and non-profit projects could increase the sustainability and impact of promoting digital democracy.

Non-profit projects can compensate for their reduced ability to shape things as passive members by being trustees of disciplines that are important to them. Here they have the opportunity to organize founding competitions and thereby exert direct influence on the selection of the active members.

—