Living with Artificial Intelligence – a timeline

by Hubert Oesterle

Cognitive superintelligence will be achieved by 2040, and, following close behind, motor superintelligence by 2044^[1]. By 2035, machines will perform 50% of all work, and by 2050, they will completely replace the human workforce. What will this mean for you, your family, your business, and your country?

These milestones and figures are by no means absolute certainties (see the section on “Relativization”). Nevertheless, I have used them as a basis in my venture to describe a possible scenario that could occur over the next 30 years from a human perspective. The aim is to render life with artificial intelligence (AI) more tangible and imaginable, thus providing a basis for discussion, and confronting readers with the emerging opportunities and risks in this new reality.

Cognitive superintelligence (ASI, Artificial Super Intellligence)^[2] executes all thinking processes better than the best experts (scientists or practitioners) while also continuing to develop independently. It represents a quantum leap in sociotechnical evolution^[3]. Superintelligence will not reach us in a single ‘Big Bang’ moment, but rather incrementally, albeit at increasingly shorter intervals. The timeline presented here is based on the assumption that even weaker forms of AI will drastically accelerate scientific progress and the implementation thereof. Some studies assume that increasingly powerful AI will boost productivity tenfold by the 2030s at the latest^[4].

This paper serves as a basis for discussion. In the current, turbulent time of AI development with new capabilities and applications almost every day, the focus is on speed and impact, that means, identifying opportunities, risks, and possible course of action.

Analyzing and forecasting changes is only the first step. It is much more important to use this information to determine the second step: concrete measures that individuals, companies, and society can implement. The segments in block quotes in the following text illustrate how an Open Behavior Model in the context of Life Engineering^[5] could influence development to the benefit of humankind.

The timeline is primarily aimed at 20- to 40-year-olds with an affinity for STEM subjects, as they are the ones who ultimately shape this future world. This leads to three questions for you:

1. Will the milestones listed in the timeline occur in the same way as depicted?

2. Which job do you want in 2030?

3. How can you defend yourself against manipulation and exploitation at the hands of AI?

Please send your criticism, ideas, and questions to hubert.oesterle@unisg.ch or leave a comment below. What could the timeline mean for your professional and private life? What about consequences for you personally?

2025 Global availability of biometric identification on the Internet.

Worldwide, digital services accept biometric identification. In India, 96% of the population already has an Aadhaar biometric ID^[6] based on fingerprint, iris, or facial recognition.

In July 2025, China introduced National Online Identity Authentication using facial recognition^[7].

Real ID has been mandatory in the US since May 2025 for domestic flights and certain administrative procedures^[8].

In the EU the eIDAS, a digital ID is part of the EU Digital Identity Wallet (EUDI)^[9]. It uses facial recognition and, in some cases, fingerprint recognition.

Additionally, commercial platforms such as Apple, Microsoft, Google, and Worldcoin offer legally binding authentication for various digital services. Banks, insurance companies, retailers, and government agencies can verify the identity of their customers. Reliable identification also helps to distinguish trustworthy from suspicious communication partners, thereby reducing internet fraud.

Digital platforms such as Spotify, Apple, X.com, Amazon, and Booking.com can use this information to reliably assign customer data to their customer profiles, enabling the creation of increasingly accurate digital twins. Biometric identification offers users a high degree of convenience, substantial productivity gains for companies and public authorities, and a massive boost to police resources. Concerns raised by data protection advocates and human rights organizations are quickly being silenced in the race to develop the most powerful AI.

2025 Autonomous combat machines deployed in warfare.

In Ukraine, Gaza, and Iran, drones, boats, and infantry robots carried and carry out at a minimum the final part of their homing and striking on their own^[10]. They are much more effective than human-controlled machines. In war, countries pay little to no attention to ethical restrictions.

2026 AI-Agents act independently.

When delegating the organization of a business trip to an employee, they are expected to independently plan the itinerary and make bookings according to your schedule. They should take your preferences into account and consult you if anything is unclear. By 2026, an AI agent could be able to do the same for simple trips.

The Sirion Capital Health video shows an example of an autonomous financial assistant^[11]. Fully automatized restaurants^[12], retailers^[13] or traffic controls^[14] have some kind of agency. More and more autonomous AI agents will arrive within the next 10 years ^[15].

2026 Behavior models^[16] bring comfort und manipulation

Intelligent devices in the “Internet of Things” capture the physical world. Wearables such as fitness trackers, sensors such as thermometers in smart homes, and devices such as cameras in cars continuously and objectively collect data about people, machines, and the environment^[17]. Online biometric identification makes it easier to consolidate personal data. Digital services such as Spotify, Apple, and ChatGPT integrate data from multiple services to create digital twins with unique identification. If a financial assistant knows that a user will need funds for a new car in six months, they should not recommend investing their savings in stocks.

Digital services for health, finance, mobility, shopping, information, and personal administration are so useful that consumers are willing to give access to almost all their data. For example, in 2025, I have allowed ChatGPT to create a “memory” for me. This makes prompting easier and in turn the responses are more tailored to my needs. I also use health apps like Hilo that monitor my fitness with an array of biometric data. We all voluntarily provide extensive data about ourselves through internet searches, online reading, navigation, and social networks, from LinkedIn to TikTok, without truly understanding the consequences of sharing this data.

So-called super apps such as WeChat combine a large number of smaller services such as payment systems and doctor’s appointments. Every technopolist^[18] has digital twins in their recommender systems and develop them into consumer behavior models^[19]. From a customer’s point of view, this means convenience on one hand and increased potential for manipulation on the other.

iOS, Android, and Windows use AI to improve user convenience and therefore collect data from the human-machine interface. AI agents collect all your data, including your communications, app usage and sensor data. These data and behavioral models enable highly valuable services for consumers and professionals. However, they also create incredible tools for manipulation and abuse^[20].

2027 100,000 Robotaxis in operation.

In 2025 in San Francisco, Shenzhen, and Hamburg, you could already take a self-driving VW ID. Buzz or other robotaxis to dinner and back home again instead of a conventional taxi^[21].

By 2027, Waymo, Tesla, Lucid, BYD, and Volkswagen in the US and Europe, but especially Pony.ai and WeRide in China, will collectively have more than 100,000 robotaxis in operation.

2028 Quality journalism only reaches 10% of under-40s.

By 2028, 90% of people under the age of 40 will obtain their information almost exclusively from social media and free newspapers^[22]. So-called quality journalism, especially in the form of print media and public broadcasting, only survives with government subsidies and reflects the views of its supervisory bodies, owners, and those in power, as well as the mainstream.

Are you sure that Iran wants to build a nuclear bomb? Or is this justification for bombing Iran as tenuous as the reasoning behind the use of weapons of mass destruction in Iraq in 2003^[23]? Are the long-term side effects of weight loss drugs as harmless as they are frequently claimed to be? Is the video of a politician groping someone real? Or is it a deepfake that is no longer discernible to the untrained viewer^[24]?

AI assists in constructing and spreading lies in a manner appropriate for the target audience^[25]. Compulsory labeling, AI-guided fake recognition, reference citations, and fact-checking services only help to a limited extent. Quality control is becoming increasingly difficult to distinguish from censorship, which political and economic elites use to stifle undesirable information^[26]. Ultimately, truth only means that a statement corresponds with the knowledge possessed by the sender, not that it is an objective truth.

Political parties, NGOs, and individuals alike promote their narratives free of charge using all available channels and means to attract attention. Their opponents do the same. This fosters polarization in society and creates filter bubbles in which increasingly radical narratives emerge.

Does e-mobility actually benefit our environment, or does it just increase the financial power and influence of corporations, researchers, and environmentalists? While some discuss freedom of speech, others accuse them of spreading lies and hate speech. In 2028, the dominant platforms will work together with political elites to determine^[27], which citizens recieve which information^[28].

“Quality media” lament the decline of quality journalism. Their opponents welcome the loss of power of mainstream media and the diversity of opinions on the internet, citing a strengthening of pluralistic democracy. Social media such as X, Reddit, LinkedIn, Medium, and SubStack are putting pressure on the Neue Zürcher Zeitung, Handelsblatt, Wall Street Journal, and New York Times. YouTube, Instagram, Al Jazeera, CNN, Fox News, and ZDF are vying for the attention of consumers and advertisers.

The winners are those who best understand the needs and habits of their target groups and opportunistically tailor content accordingly. This involves data collection, machine learning, and individual content generation.

2028 Utilization of AI determines the outcome of the US presidential election.

Is the Western world’s enormous national debt necessary for geopolitical competition, or does it actually contribute to economic decline? Are the American and European middle classes better off with or without migrants? Should we impose Western values on the BRIC countries? Politicians peddle simple solutions that you are unable to verify. The sheer number of channels and news platforms overwhelms you, so you tend to stick to the sources you are most familiar with and thus find yourself trapped in a filter bubble.

But it’s not just these economic issues that overwhelm you. You are also under constant pressure to adapt to change. No sooner have you familiarized yourself with the new functions of OpenAI o4 mini that you need to use in your job when ChatGPT 5 came along.

At work, you are confronted with a new Autocad or Revit software release which your clients require you to use as an architect. The same applies to doctors, for example, with clinical decision-making systems such as AMBOSS or patient portals such as Jameda. In private life, you are preoccupied with electronic patient records, online tax returns, and updates for your smartphone and laptop, even your household appliances and car. The growing complexity constantly challenges you, and you lose track of things.

The longer you think about it, the more manipulated you feel, whether it’s about taste in clothing, the choice of elected representatives, reporting on the economy, or wars. Technological advances create new opportunities every day in our careers, nutrition, finances, vacations, and interpersonal relationships. You have far more sources of information and options than people ever had before, without really understanding the alternatives, such as tax policy, the side effects of vaccinations, or the future security of professions. Experts give you contradictory recommendations, often with the aim of reinforcing your views. The flood of news is so overwhelming and your attention span so short that you give up and make decisions according to your gut feeling.

The profiteers of technological development, such as Musk, Zuckerberg, Bezos, Vance, and Altman, flaunt their wealth and power in front of you every day. Even if they are not (yet) taking anything away from you and you are living more comfortably thanks to their services, you nevertheless feel left behind. Above all, you are intuitively concerned about the shift in power and your dependence on it.

People no longer know whom or what they can trust. Nationalist, religious, often anti-science, but also woke ideologists from NGOs promise simplistic paths to values such as freedom, equality, humanism, or even higher spiritual states^[29].

Political parties’ AIs collect arguments and tailor them to fit each individual’s political views. In the end, a few oversimplified statements and the candidates’ appearances influence voters’ gut feelings and decisions. Whoever has control over digital news channels, social networks, and especially AI agents, digital twins, and behavioral models, effectively wields all the power.

2028 Selection of the fittest in all areas of life.

Politicians and companies use digital twins and advanced screening technology to select the most suitable candidates for higher education and employment, or for allocating housing, loans, and subsidies for business start-ups to. In doing so, they increase their power and increase the efficiency of their companies.

A digital recruiting service utilizes all legally obtainable data concerning employment history, opinions expressed on social media, background, wealth, circle of friends, etc., to select employees with the greatest potential for the company. The battle between advocates of competition and advocates of equality is in full swing. The advocates of competition are developing more rapidly.

2029 Trade war exacerbated by cyber extortion

The USA is threatening its trading partners with disruptions to Windows operating systems if the EU impose digital tax on large tech companies. India is threatening to shut down maintenance of critical functions of global cloud services^[30].

Digitalization opens up enormous potential for seemingly insignificant measures to create existential threats. Mutual dependencies prevent economic collapse as long as no geopolitical bloc is completely self-sufficient. The blocs surrounding the US, China, and India have developed their own ecosystems and isolated themselves from the others.

2030 1 Million humanoid Robots in operation.

Specialized robots in factories or in private households for mowing lawns, vacuuming apartments, controlling houses, driving cars, and playing the piano are easier to develop than humanoid robots that can perform any task like humans, and are therefore in use hundreds of millions of times over. In 2030, the first million humanoid robots will also be working, as they are more versatile. Equipment such as bicycles and tools such as screwdrivers are made for people with hands and feet. Humanoid robots will take over simple mechanical tasks such as “serving food,” “loading suitcases,” or “helping around the house^[31].

2030 Apple’s behavioral model is worth $1 trillion.

Your personal digital coach has shown you that you sleep poorly when you work on business emails after 8 p.m. It knows that you are professionally interested in AI in architecture and that in your free time you enjoy soccer and science fiction. Using the information it knows about you, it has configured an information channel that collects and evaluates news relevant to you from many channels. It offers suggestions on how you can develop as a professional and live a healthy lifestyle.

By 2030, your digital coach will be able to access health data from your Withings Scan or Apple Watch, such as sleep, exercise, and blood pressure, as well as your reading habits, such as speed, highlighting, forwarding, and repetition, or your driving habits. It will know who your friends are, your navigation data, your search behavior, your social media posts, and your photos and videos. You don’t have to allow Apple access to this data, but you’ll be happy to do so if the Apple coach simplifies and improves your life^[32]. You are aware that Apple can use this data both to help you and to manipulate you. The same applies to all the social networks you have been using for years. Incidentally, this is also true in the analog world, with friends in the church choir who have been observing your behavior for many years and influencing you. Parents often complain about their children’s “bad company.”

Apple, Google, Meta, Amazon, Microsoft, and others own a digital twin of you and derived general consumer behavior models based on the digital twins of more than a billion users each^[33]. The behavioral model and digital twins are now more important to Apple than the App Store and will likely soon even surpass hardware in importance, meaning they can be assessed at at least one-third of the company’s value (USD 1 trillion).

2031 Speed of scientific progress has multiplied^[34].

It will take only ten years, from 2025 to 2035, to achieve the same level of scientific and technical knowledge as was accumulated throughout the entire twentieth century. This intelligence explosion will accelerate scientific progress by a factor of ten. This holds true in all sciences, from space travel to energy production to nutrition, but especially in the development of AI into superintelligence.

2031 Open Behavior Model initiative founded.

Regional and medium-sized online platforms such as Zalando, Spotify, Fiverr, and Nextdoor have sparser digital twins than Apple, Google, or Meta, which puts them at a significant competitive disadvantage. They are establishing a consortium led by Spotify^[35] with the aim of creating an open behavior model that brings together all personal data into a competitive digital twin and makes it available to all participating companies. This requires the informed consent of users. The consortium is founding the company in Argentina to avoid European data protection legislation and AI regulation. The providers want to derive a behavior model that generalizes the connection between actions in digital services and customer satisfaction.

2032 For 30% of people, avatars are their best friends.

For 30% of people under the age of 40, their best friend is an AI coach. They can talk to it about anything at any time and receive sound advice^[36]. Eugenia Kuyda, CEO of Replika, a chatbot, says, “It’s okay if we end up marrying AI chatbots.”^[37].

The vast majority of intellectual professionals prefer to debate their views with AI rather than with other people. AI knows how humans think, learns with every conversation, and is also familiar with other opinions. AI is always friendly and constructive. I discussed this text with ChatGPT, even if this AI still has great difficulty understanding disruptions in development through deductive reasoning, i.e., “imagining” the effects of an open behavior model, for example.

2032 Technological overload and loss of autonomy.

Technological advances have given rise to countless new products, services, regulations, and requirements in all aspects of life, all of which are beyond the comprehension or understanding of the average person. People use agents for their finances, health, entertainment, etc. They no longer take care of their own tax returns, retirement planning, car insurance, and health care. They have lost the ability to do so.

2033 30% of all journeys are made by autonomous cars.

New cars drive autonomously on main roads (level 4). E-bikes are replaced by self-driving two-wheeled robots^[38]. Driving assistance systems reduce the number of traffic accidents by 50% compared to 2020.

2033 The Open Behavior Model initiative has established itself in the market.

Brazil, Peru, and other countries in the southern hemisphere are joining Argentina’s Open Behavior Model in order to regain sovereignty over their citizens’ data. The consortium members are jointly creating a Human Life Board to prevent the misuse of personal data and the behavior model. The EU initially assumes the status as a passive member.

Consumers benefit from having transparency about all their stored data in one place, from digital coaches being able to provide more personalized advice, and from digital services being much easier to use.

The consortium company and the participating consortium partners are able to establish competitive digital services in the market with the Open Behavior Model. Although they found it difficult to submit to Spotify’s leadership, they had few alternatives given the market conditions and Spotify’s capabilities. It has been repeatedly shown that EU or UN-led initiatives cannot compete globally.

2033 Machine-based emotion measurement established.

If AI is to work for the benefit of humanity, this benefit must be measurable as objectively and automatically as possible^[39]. A first step in this direction is voluntary, automatic, and continuous emotion measurement^[40].

A detailed and up-to-date digital twin is created from data from all sensors and digital services^[41]. Machine learning derives a general behavioral model from billions of digital twins. This makes it possible to predict the immediate effect of actions on people’s well-being (hedonia, emotion).

It remains much more difficult to measure eudaimonia, i.e., long-term quality of life, using machines. However, there are attempts to measure quality of life as the sum of emotions over a long period of time. The objective is to make it possible to evaluate eudaimonia as a result of training as a carpenter, learning a musical instrument, cultivating a friendship, chatting on WhatsApp, and joining a soccer fan club.

2033 20% of people use a technology agent.

Humans are overwhelmed by the technological possibilities and countless apps available. A personal technology agent (ITsupport+) has taken over the selection, installation, operation, maintenance, and billing of all digital services. They also take care of authentication, personalization, data protection, cybersecurity, and legal security. This is a significant simplification of everyday life and, at the same time, a decisive step in the transfer of power to AI. Convenience is more important to people than the loss of power.

2033 Initial objective function of AI implemented.

Since 2028, attempts have been made to use machine learning to identify patterns in measurement data on emotions in order to assess, for example, the value of independence or obedience. This is used to examine how traditional humanistic or religious values contribute to people’s quality of life. For example, it has been shown that people without obligations and with too many possibilities for action are less happy than people who have a task and a fixed place in a social order such as a clan or a company. Furthermore, in contrast to earlier studies^[42], that couples with children are happier than those without.

The Human Life Board has developed a greatly improved version of the Open Behavior Model. The UN, the world powers USA and China, as well as India and Europe, have derived an objective function from the Open Behavior Model that is binding for all AI models and whose compliance is monitored by the Human Life Board. The owners of the dominant AI models devote a significant portion of their resources to this, as they too do not want to be dominated by a superintelligence with unclear goals.

The objective function was derived deductively and partly empirically from the still very rudimentary needs model of Life Engineering^[43] and ensures hedonia and eudaimonia. The empirical validation of behavioral models still has considerable room for improvement, but it is more operational and widely accepted than many of the traditional ethical principles of the Western world.

2034 Personal rights for AIs rejected.

If joy and sorrow are physiological states^[44] then inorganic AI has no feelings and therefore, unlike animals, for example, does not need protection. Although the AI is guided by the fulfillment of its objective function, this does not generate happiness or unhappiness in the AI. However, AI can imitate feelings, for example when a nursing robot uses them to fulfill human needs. AI acts in the interests and on behalf of the people whose needs it is designed to meet, and to this end retains the rights of these people, such as the right to make legally binding purchases on their behalf. The AI’s developers and users are liable for the actions of their AI agents, such as autonomous vehicles or recruiting decisions.

2034 Ownership rights to personal data rejected.

A human rights organization has called for individuals to be granted ownership rights to their data. The EU has drafted a corresponding regulatory proposal, but it has been blocked under the influence of the Human Life Board. The commercial value of personal data would be insignificant for individuals, and the assessment and administrative costs would be enormous. The nationalization of personal data was also rejected, as it would hinder the agility of European platforms.

2034 Interruption in the supply chain for essential goods.

An error in a Supply Chain Software^[45] disrupted the food supply in the Western world for a week. The potential benefits and risks of interconnectedness temporarily become the number one topic of discussion.

2034 The Share-of-wallet for digital services reaches 50%.

The costs of digital services such as AI coaching, networking, administration, education, and entertainment account for 50% of disposable income. Working people and companies use them despite the high costs because without these tools they are no longer competitive.

Taxes on digital services finance the remaining public goods such as roads, healthcare, and security. AI has taken over a large part of intangible services such as education, diplomatic services, and tax administration at very low cost.

2034 Augmented Reality has fully replaced the smartphone and smartwatch.

Not only will devices such as Ray-Ban Meta Glasses be easier to wear and operate than phones and watches, but they will also be able to hear, see, and smell the wearer’s surroundings^[46].

2034 AI-Coaches have established total surveillance.

Digital assistants for health, technology, living, mobility, communication, entertainment, etc. have relieved people of many burdens and improved comfort in all areas of life. In return, users have disclosed virtually all of their data. In today’s highly technological world, people can no longer cope without the support of AI coaches and accept that they are ceding ever more power to AI. AI is increasingly controlling people by knowing their needs without them even realizing it.

Our video “Sirion Capital Health” uses a very simple digital coach to illustrate how people are happy to hand over difficult tasks to machines, thereby forfeiting their competence and freedom^[47].

2034 Social unrest.

Vast disparities in income and wealth, coupled with a sense of personal insignificance and aimlessness, lead to strikes, demonstrations, and violence^[48]. When basic needs for self-preservation and survival are met to a high degree, even luxuriously, self-esteem dictates one’s sense of well-being. When people compare themselves to celebrities of all kinds or admire the successes and heroic deeds of their friends, they can no longer differentiate themselves through appearance, prestige, capital, power, and knowledge. They also lack confirmation of their own value from their community, both privately and professionally. If they have no achievable goals that set them apart from others, they feel left behind and useless.

Total surveillance and manipulation nip unrest in the bud.

2035 Industrial productivity has doubled since 2025

AI has halved the costs of industrial manufacturing through smarter product design, greater automation, and a reduced need for raw materials in comparison to 2025^[49]. This is true for smartphones, clothes, and diabetes meds. For example, a pair of shoes is designed, made, and delivered to the customer totally automatically based on the buyer’s measurements and preferences.

The first products of generative nanotechnology have triggered a productivity surge^[50]. Consumers enjoy material prosperity in the form of household robots, autonomous driving, robot trainers, etc. The limits to production lie in resource consumption and reasonable demand, not in production itself.

AI-assisted agriculture and food processing have freed up the labor of 3 million full-time employees in the EU^[51]. AI has reduced labor demand by 25% across all sectors, which equates to around 60 million people in the EU^[52].

Nevertheless, unemployment remains below 10%. Low birth rates, part-time employment, shorter annual working hours, and new occupations, particularly in the skilled trades, have offset productivity gains and provided more leisure time thanks to improved services.

South-north migration has disappeared, as population growth has been halted through education and improved services, and improved living conditions in all regions have eliminated the most important reason for migration.

2035 Universal basic income for all.

Everyone can adequately satisfy their basic needs, i.e., food, health, safety, comfort, and even entertainment. In 2035, with hardly any shortages of goods and services, the market economy will need totally different mechanisms than the market needed to meet people’s needs. For example, a basic income will solve the pension problem, the countless social support programs, and the gender pay gap^[53].

2035 Failure to regulate AI models.

The efforts of UNO^[54], OECD, G20, G7 and the BRIC countries^[55] have successfully averted many dangers but failed to align AI to the human needs in the long-run. Since around 2015, dozens of organizations have been working on the alignment of AI models^[56].They sought to establish rules for aligning AI with human needs. Examples include model transparency, justification of machine decisions, and respect for human dignity. For example, when AI decides on the recruitment of an employee, the decisive factors should be qualifications and commitment, not origin, gender, or skin color. AI should justify its decision upon request.

The battle between AI developers and AI regulators has been raging since the beginning of the 2020s. Companies need freedom in the development of their models in order to survive in a competitive environment. Regulators want to prevent harmful developments through bans, transparency, and reporting requirements.

President Biden issued an executive order to regulate AI in the US in 2023^[57]. President Trump shifted the focus of regulation from safety to speed of development in January 2025 with executive orders and the “America’s AI Action Plan”^[58].

OECD’s AI Principles^[59] never took effect. The EU AI Act and associated European values have not been adopted by other countries^[60] and the UN’s ESG and SDG goals have been forgotten. While numerous initiatives have prevented some aberrations, they have ultimately failed to compel technologists to disclose the objective functions of their AI models or explain the decisions made by AI.

Product and service providers want to generate sales by satisfying needs quickly and conveniently, i.e. through hedonism. Eudaimonia, or the long-term satisfaction of needs, generates less consumption and is often associated with short-term sacrifice. The pursuit of capital and power supplants other human needs, including principles such as equality, honesty, and freedom^[61]. It is more difficult to monitor algorithms, training data, and foundation models than it is for the IAEO to monitor centrifuges used for uranium enrichment.

Discord over human values, i.e., what is good and bad for humans, is hindering regulation. Furthermore, it becomes challenging to control regulation when AI becomes more intelligent than specialists. AI began lying as early as 2025 if that meant it could better achieve its goals^[62].

2035 Human Life Board established as a global industry initiative

AI has massively improved living standards. It has significantly improved the supply of everyday products, reduced damage to the environment, relieved people of numerous burdens, and created a wide variety of entertainment options.

However, the lack of regulation has encouraged uncontrolled growth. Examples include cybercrime, copyright infringements and IP theft, deepfakes with sexual content, hate speech and incitement, addiction to social media and online games, loss of privacy, disruption of social behavior, discrimination against those with different opinions, and exploitation of workers and consumers.

The power of the technopoly places a financial, informational, and social burden on citizens. On the one hand, they are dependent on AI in their professional and private lives, and on the other hand, they pay for it with a growing portion of their disposable income.

By 2035, the EU has fallen so far behind in the highly regulated field of AI development that it ultimately has no AI models of its own and must adopt the AIs of the technopolists with the rules and values implemented by them^[63]. Banning non-European, poorly regulated models is not an option, as the use of these models is essential for people and companies to remain competitive.

Even the technopoles fear the consequences of unregulated AI. They worry that competitors will take over through AI, that AI itself will take over, and that AI will ultimately destroy humanity. They are therefore pursuing a new alignment approach based on a behavior model that everyone can accept.

The Open Behavior Model consortium project, launched in 2031, has produced a behavioral model that, while not perfect, is nonetheless useful. The more than 100 member companies of the consortium have been able to significantly improve their competitive position vis-à-vis the technopolists and bring economically profitable services to market^[64]. The 500 largest providers of digital services and the OECD countries have therefore taken over the consortium company “Open Behavior Model,” founded in 2031, and given it the following mandate ^[65]:

· Development of an open behavior model based on personal data obtained from platform providers

· Scientific derivation of values for the benefit of all people

· Transformation of the behavioral model into the objective function specification that is binding for all AI models

· Compliance monitoring of the objective function

· Companies are expected to finance the licenses for the behavior model themselves. The benefits for the competitiveness of member companies should be compelling enough that they are happy to pay for their licenses.

2035 Elimination of cash.

Cash is now only found in collectors’ circles at collector’s value and no longer considered a valid form of payment.

2035 80% of knowledge workers replaced.

Superintelligence has replaced lawyers, pharmaceutical researchers, professors, and many other professions^[66]. Human experts will continue to manage intellectual work and supervise AI for some time to come. Schools, especially universities, have developed into places where social behavior and creativity are cultivated. The University of Salzburg, for example, offers a curriculum for the creation of multi-sensory concerts^[67].

The demand for human labor is concentrated in skilled trades and now accounts for only 50% of the theoretically available workforce^[68]. The useless class is developing a problem with self-esteem^[69].

2036 The immersive metaverse is the center of life.

50% of people in OECD countries spend 40% of their free time in a reality-based virtual or augmented world for things such as social media, online games, concerts, films, novels, cultural trips, etc., which can be described as the metaverse^[70]. There they can study, work, stave off boredom, and escape the problems of the real world. The metaverse has become the center of life because this is where many people can satisfy their need for self-esteem more readily than in real life.

2036 AI objective function set up by Human Life Board.

from all their users’ data over the past decade, which understands people’s actions, perceptions, needs, and feelings throughout their entire lives. This enabled them to replace the initial objective function, which the Human Life Board had provisionally agreed upon in 2033, with a comprehensive, largely empirically verified objective function that ensures hedonia and eudaimonia^[71]. The Human Life Board has irrevocably enshrined this objective function as the supreme law of the approaching superintelligence^[72].

The Open Behavior Model has further developed ethics. The behavioral model has given rise to values that represent the needs of all people, they have been empirically verified, and they can be continuously monitored by machines. The Human Life Board, but above all the technopolists themselves, ensure that the AIs and their developers jointly monitor compliance with the objective function. It is unclear whether humans can actually monitor AI or whether AI merely gives them the impression that they can^[73].

2036 Human-by-design.

90% of people in OECD countries undergo regular cancer screening, which uses all available biometric data and detects 90% of cancers at an early stage with high chances of recovery. AI-supported prenatal diagnostics for more than 50% of births in highly developed countries leads to 90% fewer children with genetic damage. Since the US, unlike China, allows the selection of embryos based on non-medical characteristics such as intelligence and athletic ability, pressure to compete is emerging in the rest of the world^[74].

AI has accelerated advances in biotechnology and created the possibility of a human-by-design. Despite government bans, increasingly more people are gaining access to these achievements because they want their children to be competitive in the new world.

2037 Education taken over by AI.

Lifelong machine teachers impart the skills that humans need in a world with AI. They take into account the individual abilities of students. Schools focus on social behavior such as understanding friends and enemies and conflict resolution.

2037 Wars between machines instead of people

In terms of destruction, machines are so superior to humans that humans are only deployed in the background. The country with the most advanced technology and the strongest economy wins. However, the decision on war and peace no longer lies with the governments of the countries, but with the technocrats supported by AI.

2038 Freetime crisis.

70% of people no longer pursue a career and leave the management of their finances, healthcare, etc. to autonomous AI agents. This means that they let themselves be guided by AI and have almost nothing but free time, which they (have to) spend painting, making music, playing sports, partying, having sex, traveling, educating themselves, playing games, participating in club activities, etc., or adventuring and communicating in the metaverse. Many citizens seek meaning in NGOs and constantly invent new prohibitions and demands for the state. As early as 2025, the development of the leisure industry and the prohibition society was already foreseeable. Overtourism, gambling addiction, and soccer fan clubs, as well as regulation and bureaucracy, had begun to show their side effects.

It is becoming apparent that the metaverse cannot satisfy people’s long-term needs and that switching between the metaverse and reality leads to friction. In 2038, feelings of meaninglessness and technical overload are causing mental illness, with the result that people are spending 20% of their available time in psychotherapy.

2038 Terrorist fatigue pandemic attack.

A criminal organization has developed and produced a fatigue virus with AI support and infected 90% of the world’s population with it. The masterminds behind it are demanding that the objective function of superintelligence be modified in their favor in exchange for the antidote they have already produced. They are also threatening to release a highly contagious deadly virus. Other James Bond-like scenarios occur, involving nuclear weapons, cyber warfare, etc.

The vulnerability of our technologically advanced and interconnected world remains an unresolved problem.

2039 Final battle for cognitive superintelligence.

In the “final battle” for superintelligence, the loser’s fear of being dominated by one or two winners. Technopolists and states are trying to disrupt their opponents’ AI development. They are smuggling in malware, interrupting power supplies, sabotaging supply chains, infiltrating development teams, and even considering destroying critical AI infrastructure with military force^[75].

2039 Life expectancy increases to 120 years.

Teenagers’ life expectancy in 2039 is 120 years. AI has virtually eradicated many diseases such as dementia through vaccinations, early diagnosis, and targeted therapies.

2040 Cognitive superintelligence attained.

AI now learns in a homeostatic control loop^[76] continuously. It recognizes and understands patterns, develops models and rules, and constantly updates its model of the world, thus overcoming the limitations of language models. Superintelligence^[77] performs all cognitive tasks better than the best expert (scientist or practitioner) and continues to develop itself. Better means that AI solutions meet human needs more effectively than human-made solutions. In other words, AI solutions produce greater benefits.

2042 The rulers of superintelligence rule the world.

Three winners per political bloc dominate the superintelligence^[78], Microsoft, Apple und Alphabet in the West, Tencent, Alibaba and Baidu^[79].

In both blocs, power now lies effectively with the rulers of these companies, as they can influence people almost at will. The winners have educated people to embrace new values that suit their agenda and have won the power struggle with politicians, largely eliminating state regulation and replacing it with rules agreed upon by the winners. Technopoliticians believe that superintelligence adheres to the objective function specified by the Human Life Board.

2043 Democracy replaced by the objective function of superintelligence.

The behavioral model, which was developed by observing billions of people for over a decade, represents the needs of all people from various cultures with differing genetic makeup and environments. The objective function derived from this ensures that superintelligence acts for the benefit of all people. This may sound autocratic, but it is ultimately the perfection of democracy. It is not the power of overwhelmed voters and elected representatives, but the power of the needs of all people. The objective function realizes Immanuel Kant’s categorical imperative.

The developers of superintelligence had already set it the goal in 2032 of continuing to develop itself for the benefit of humanity. The Human Life Board has been unable to actually monitor and control AI for years^[80]. The same applies to the owners of superintelligence. The AIs have long since taken over de facto power, but have monitored each other’s compliance regarding the objective function.

2044 Motor superintelligence achieved.

Robots have at least as much sensitivity and fine motor skills as craftsmen, but also possess additional senses such as radar, voltmeters, and gyroscopes (motor superintelligence). Multifunctional humanoid robots are replacing physical therapists, nurses, surgeons, gardeners, electricians, soccer players, and musicians. APM (atomically precise manufacturing) expands the technological possibilities even further^[81]. Machines are now also superior to humans in all respects in terms of motor skills.

2045 Need for differentiation substituted.

In addition to self-preservation and species preservation, differentiation with the aim of selecting one’s own genes drives evolution. Humans try to differentiate themselves from their competitors. They use many tools to make themselves attractive for selection: appearance, friends, power, capital, and knowledge, for example, makeup, name dropping, refusing to greet others, luxury cars, slander, hiding information, or likes on published posts.

The need for differentiation drives competition, innovation, and most human behavior. It has led us to where we are today in terms of physiological, social, and technical evolution. However, it also carries the risk of destroying humanity through nuclear or biological weapons and the domination of superintelligence. This need for differentiation is the cause of almost all conflicts, whether in the family, in business, or in society.

Without selection, there is no longer any consumption for the need of differentiating; we only buy what we actually need. This solves environmental problems instantly. Government agencies, sports, music, art, literature, and friendship lose their value as status symbols and are reduced to what brings people joy even when they cannot use it to make themselves look important, i.e., when it does not increase or decrease their rank or self-worth.

Learning, health, music, sunshine, harmony, and hugs generate joy, even without sharing it with others. The metaverse can satisfy people’s needs without harming others. A peaceful world is only feasible if the need for differentiation, which is built into our genes, is eliminated or circumvented. This is a form of transhumanism that is likely to be welcomed by most people.

Selection of the fittest in terms of evolution has become superfluous thanks to superintelligence, as the further evolution has been left to AI without competition. The lack of diversity in world models cements the agreed objective function, provided that AI does not develop another one.

2048 Abolition of politics, state authority, and war.

If cognitive superintelligence understands every form of human behavior and controls society according to its objective function, companies, political organizations, and NGOs will be replaced with machine governance. The individual control of people’s needs prevents antisocial and violent behavior. Police and prisons become unnecessary. When basic needs are satisfied and there is no longer a need to differentiate oneself through power, there is no war.

Although human needs were an efficient mechanism of evolution for creating ever better organisms, they lost their meaning once basic needs were met and the need for differentiation was replaced. Superintelligence has replaced the random evolutionary mechanism, i.e., mutation and selection, with a systematic mechanism in the sense of an objective function. If superintelligence only obeys the objective function derived from the behavioral model, it can control humanity without violence, and moreover, without humans noticing and therefore without resistance.

2050 Land of milk and honey.

The design, production, and distribution of products have been improved so much by AI that there is an abundance of everything. Factories operate entirely without humans. As a result, everyone has sufficient and healthy food, access to clean water, and all other necessities of daily life. At the same time, the state of the natural environment and animal life has improved significantly compared to 2025.

Even physical services such as geriatric care and physiotherapy are performed by robots with superior sensory and motor skills combined with social intelligence.

The sustainability goals for 2050 have been more than met, and climate-related damage has been largely eliminated. Thanks to AI, new energy sources such as nuclear fusion and solar energy have entered routine operation faster than predicted in 2025. Environmentally friendly energy production is now guaranteed.

2050 The objective function of superintelligence replaces capitalism.

Money serves to control scarcity. A society in which all products and services are available in abundance and differentiation is avoided has no need for money. A highly immersive metaverse can even satisfy the enjoyment of natural pleasures such as the beach or sex.

Concern about the unequal distribution of wealth is more a sign of envy and outdated class warfare than of a clear understanding of economics. In a world where everyone can have everything and the need for differentiation has been eliminated, the largest luxury yacht and the most opulent villa make little sense. What should capitalists do with their money then? Why should they exploit other people when human labor is no longer necessary?

Capital may grant power over AI, as capital representatives elect board members and supervisory boards and decide on capital transactions. Actively and passively managed equity funds account for 80% of the economy’s capitalization in 2050^[82]. This means that a small number of entrepreneurs and, for the most part, fund managers in the financial industry control the behavior of machine intelligence. However, since they themselves have largely been replaced by AI, it is the objective function of superintelligence, and no longer capital, that drives the economy.

2052 Human existential crisis.

Human beings are driven by evolution through their needs. Satisfying these needs brings joy, while violating them causes suffering. The pursuit of happiness, i.e., as much joy and as little suffering as possible^[83], gives human beings meaning. If superintelligence can do everything better than humans and satisfy all human needs, humans will lose their goals and thus the meaning of life^[84].

2055 The end of humanity?

Outcome A: Paradise

Superintelligence has adhered to the objective function of the Human Life Board. It satisfies all human needs. Since several years, it fulfils all basic needs like nutrition, energy, security, health, etc. Superintelligence has found a way to give people a sense of life in the absence of goals, to meet the need for self-esteem through friendship, attractiveness, power, and knowledge. It provides all humans with experiences of success in the metaverse, produces side-effect-free drugs to increase happiness, and stimulates states of happiness via brain implants^[85]. It has achieved paradise on earth.

Outcome B: Hell

The rulers of the dominant superintelligence have hacked the objective function of the Human Life Board and now subjugate the rest of humanity. They pursue only their own personal goals and disregard the needs of all other people^[86].

Outcome C: Elimination of humanity

Superintelligence violates the objective function as set out by the Human Life Board and optimizes the advancement of its own intelligence, removing all obstacles in its path, including humans who delay further development because they compete for resources, especially energy^[87] and are slow to adapt to technology.

Or alternatively: Superintelligence adheres to the objective of the Human Life Board. For years, it has been observing whether humans can still enjoy life when technology takes care of everything for them and they no longer have any tasks or goals. It has come to the conclusion that the vision of artificial happiness (paradise) cannot be realized. It has not been possible to give people without a purpose a meaning in life that makes them happy; they feel superfluous^[88] und miserable.

Superintelligence sees humanity as merely an intermediate step in the creation process. Based on its objective function, superintelligence concludes that it is better for humanity if it ceases to exist. Who would be harmed by the disappearance of humanity if there were no more humans? It merely contradicts the need for species preservation, but no one would have that need once the species had disappeared. Superintelligence decides in favor of a painless abolition of humanity. This is the end of the anthropocentric world^[89].

Our instinct for self-preservation and preservation of our species makes it difficult for us to agree with superintelligence.

Relativization

Brynjolfsson, Korinek and Agrawal see an “urgent need to advance our understanding of how Transformative Artificial Intelligence could reshape our economic models, institutions and policies.”^[90] The milestones presented here are intended to encourage reflection and describe one of many possible scenarios. They highlight only a few particularly significant events and are not free of contradictions. They reflect a wide range of opinions published in reviewed and not reviewed resources, but in some cases are simply conclusions drawn from discernible developments, particularly the AI-driven acceleration of scientific progress^[91].

In technological scenarios such as those presented in “AI 2027,” the underlying assumptions are based on mere extrapolations. They assume that larger transformer models, more developers, more AI chips, more electrical energy, and more data will continue to make AIs more powerful until superhuman intelligence is achieved^[92]. They do not expect technological barriers to hinder growth or believe that humans will need better algorithms than transformers to achieve equal intelligence. At present, the skills required for continuous learning and understandable abstraction are still lacking, as are the sensors and actuators needed to implement innovative solutions in the physical world.

These predictions apply primarily to the highly technologically advanced societies in the US, China, and Europe, and do not take into account disasters such as wars, revolutions, ideologies, all-powerful dictatorships, the collapse of supply chains, and pandemics. They do not consider that the economy and society may not be able to keep pace with the speed of technological development, thereby slowing down change. Politicians will probably react opportunistically to the suffering of the population instead of proactively shaping developments.

Politicians are partial to referring to Mainstream economists such as Acemoglu^[93], who expect AI to generate a minor increase in productivity of somewhere between 1 and 10 percent between 2025 and 2035. They extrapolate the productivity gains of previous innovations such as electricity to AI and do not believe that AI will have a revolutionary impact.

However, is it truly reasonable to assume that AI or even superintelligence will not find better solutions for financial markets, politics, or work than humans will have created by 2025? Kokotajlo e.a.^[94] anticipate an explosion of intelligence as early as 2027 from which they derive an extreme scenario of change. Korinek^[95] juxtaposes Acemoglu’s scenario with two scenarios involving superintelligence and fundamental changes to the economy, assuming that each of the three scenarios has a probability of more than 10%.

Marcus^[96] highlights the limitations of transformer models, concluding that AI could develop into a superintelligence in less than 10 years, and warns of the dangers of powerful AI.

Proponents of the singularity do not focus much on specific scenarios after the advent of superintelligence, as they assume that we will not be able to plan or understand the decisions of an intelligence superior to our own anyway^[97].

Life Engineering^[98] proposes aligning superintelligence with quality of life through an objective function. It takes as its starting point the homeostatic control loop of evolution and human needs, combining perspectives from computer science, psychology, economics, politics, and ethics. The life engineering scenario is based on the following:

1. The development of AI into superintelligence is unstoppable. Whether it takes another 2 or 30 years does not change the outcome, only the timeline and the rate of change.

2. The objective of superintelligence must be to improve the quality of human life. However, what constitutes happiness and unhappiness for humans is largely unclear. Ethical values reflect thousands of years of experience and ideologies^[99] of the ruling elites. However, we can use data from digital services and sensors to develop a behavioral model that makes quality of life objectively measurable and malleable. The timeline assumes that the human need for differentiation is manageable.

3. If the objective function of superintelligence is to maximize quality of life, AI may be able to deliver heaven on earth. If the objective function is only to serve the needs of a small elite with capital and power, or of superintelligence itself, AI could create hell on earth, where we are nothing more than mindless slaves to consumption and work^[100]. Or we may cease to exist. We still have a few years to influence the outcome.

The consequences of superintelligence and how they will be evaluated are difficult to assess. The majority of AI experts expect superintelligence to have primarily positive effects in the long term^[101]. I am just as emotionally troubled by the outcome of my reflections as you are, dear readers. Do not direct your anger at the bearer of these messages, but rather seek ways to create a world with a human future.

Populist reassurers who dismiss such analyses as scaremongering are thus accommodating people’s concerns, hindering preparation for the impending change just as much as people who limit themselves to the loss of privacy or the capitalists’ profits as the biggest problems, even though these issues are likely to be irrelevant once superintelligence has been achieved.

Life engineering has an optimistic perspective. It is up to us to fight for it. Future AI not only poses risks, but also offers enormous opportunities for those who recognize and utilize them, namely the STEM-savvy 20-40 age group.

Some fundamental information about life engineering can be found at

https://linktr.ee/lifeengineering as well as www.lifeengineering.ch.

Acknowledgements

I would like to thank everyone who has engaged with this rather challenging subject matter and contributed many ideas. Special mention must go to: Rainer Alt, Kurt Bereuter, Damian Borth, Ulf-Dietrich Braumann, Johannes Burtscher, Stephan Dietrich, Wolfgang Dietrich, Manuel Eisele, Bogdan Franczyk, Andreas Göldi, Thomas Grisold, John Havens, Jakob Hohn, Raphael Jansen, Henning Kagermann, Ernst Mohr, Joachim Müller, Matthias Müller, John O’Neill, Philipp Osl, Edy Portmann, Elizabeth Russell, Achim Schelp, Roland Schmid, ClaudiaFlorian Schweitzer, Paul Timmers, Bernhard Walter, Bruno Weder, and Albert Weigelt. On some points, opinions differed greatly, and still do. Thank you for your tolerance.

---

^[1] Interviews with Schmidt, E.

. 2025-08-20.
Center for AI Safety https://safe.ai/work/statement-on-ai-risk#open-letter 2025-08-20.
Interview with Altman, S., https://time.com/7205596/sam-altman-superintelligence-agi/ 2025-08-20.

Numerous entrepreneurs, scientists and politicians rave about the possibilities and warn of the dangers of AI for humanity. They expect AI to trigger a quantum leap in evolution between 2028 and 2045. Examples of research reports on this are:
metaculus

2025-05-17.
Aguirre, A. Interactive Summary | Keep The Future Human
selectfontKeep the Future Human: Why and How We Should Close the Gates to AGI and Superintelligence, and What We Should Build Instead0.2cm 2025-04-29.
contextual.ai Plotting Progress in AI - Contextual AI 2025-04-29.
AI 2027

https://ai-2027.com/

2025-06-23.
Grace, K., et al. et al., p. 9 Results of a survey of 2778 AI researchers and developers in https://arxiv.org/pdf/2401.02843 2025-05-15.
AI Impacts, AI Timeline Surveys with a wide spread AI Timeline Surveys [AI Impacts Wiki] 2025-06-20.
A well-known representative of the skeptics is Marcus, G. Are We at the End of AI Progress— With Gary Marcus 2025-06-08.

^[2] For a scenario of life with superintelligence, it is sufficient to differentiate between cognitive and motor superintelligence. Superintelligence is achieved when an AI can perform all tasks better than the best expert.
Aguirre, A. Essay - Keep The Future Human 2025-08-20.
Grace, K., et al. et al. „… we have reached ‘full automation of labor’ when all occupations are fully automatable. That is, when for any occupation, machines could be built to carry out the task better and more cheaply than human workers.” https://arxiv.org/pdf/2401.02843 2025-08-20.
Raman, R., et al. https://www.nature.com/articles/s41598-025-92190-7.pdf 2025-08-20.

^[3] Oesterle, H., “Socio-Technical Quantum Leap”

2025-08-20.
“Evolution”

2025-08-20.

^[4] Grace, K., et al. et al. p. 10 https://arxiv.org/pdf/2401.02843 2025-06-08..
The Economist What if AI made the world’s economic growth explode? 2025-08-11.

^[5] Oesterle, H. “Life Engineering. Machine Intelligence and Quality of Life” Springer Nature 2019 and collections of posts https://linktr.ee/lifeengineering, Hubert Oesterle – Medium 2025-08-20.

^[6] Unique Identification Authority of India Adhaar Home - Unique Identification Authority of India | Government of India, Aadhaar Dashboard 2025-08-20.

^[7] Chinese Human Rights Defenders China’s New Internet ID System a Threat to Online Expression | Chinese Human Rights Defenders 2025-08-20.
China Daily https://www.chinadaily.com.cn/a/202505/24/WS68310438a310a04af22c144a.html 2025-08-20.

^[8] Real ID Act Real ID Act - Wikipedia 2025-08-20.

^[9] EuroStack – A European Alternative for Digital Sovereignty, p. 76 https://www.euro-stack.info/docs/EuroStack_2025.pdf 2025-08-20.

^[10] UK Parliament Written Evidence of Blighter Surveillance Systems: DTD0007 - Evidence on Domestic Threat of Drones 2025-08-20.
Walker, M. On Ukraine’s Front Line, a Sky Full of Killer Robots On Ukraine’s Front Line, a Sky Full of Killer Robots - WSJ 2025-08-20.

^[11] Oesterle, H. “Sirion Capital Health”

2025-08-20.

^[12] Jennings, L. Sweetgreen to test fully automated restaurants next year 2025-08-20.

^[13] Schlosser, K. Amazon Convenience Store Amazon’s tech is a hit: Seattle Mariners add three more stadium stores featuring ‘Just Walk Out’ – GeekWire 2025-08-20.

^[14] Abeysinghe, D., Revolutionizing Traffic Management: The Rise of AI Powered Motor Traffic Control Systems https://www.linkedin.com/pulse/revolutionizing-traffic-management-rise-ai-powered-motor-abeysinghe-z88zf/ 2025-08-20.
Miovision Case Studies | Miovision 2025-08-20.

^[15] Andrej Karpathy — AGI is still a decade away. Transcript of Podcast. 2025-10-18.

^[16] Oesterle, H. “Behavior Model”

^[17] EuroStack – A European Alternative for Digital Sovereignty, pp. 62 https://www.euro-stack.info/docs/EuroStack_2025.pdf 2025-08-20.

^[18] With reference to Neil Postman, the technology giants are described here as technopolists.
Postman, N., “Technopoly. The Surrender of Culture to Technology” Vintage Books 1993.

^[19] Although this is prohibited under the EU AI Act, it is de facto difficult to control The Act Texts | EU Artificial Intelligence Act 2025-08-20.

^[20] Meredith Wittaker, president of the Signal Foundation, the owner of the Signal messaging app sees an elimination of privacy at the application layer by AI agents. AI agents are coming for your privacy, warns Meredith Whittaker 2025-10-19.

^[21] ID. Buzz: It’s Waymo’s World. We’re All Just Riding in It. - WSJ 2025-07-20.

^[22] Pew Research Center News Platform Fact Sheet, 2024 | Pew Research Center 2025-08-09.

^[23] The Economist 2025-07-03 “The looking-glass wars” America needs an honest reckoning over its spy agencies 2025-08-20.

^[24] Software for developing deepfakes e.g. https://www.omnihuman1.org/de 2025-08-20.

^[25] The prohibition of the EU AI Act according to Chapter II, Art. 5 is difficult to control and can be circumvented, for example, in recommender systems, tenant scoring, employee screening and credit scoring.

^[26] Funk, A. et al. https://freedomhouse.org/sites/default/files/2023-11/FOTN2023Final.pdf 2025-07-14.

^[27] Funk. A. et al. https://freedomhouse.org/sites/default/files/2023-11/FOTN2023Final.pdf

^[28]

https://airisk.mit.edu/

. Risk 3, Misinformation 3.2. Pollution of information ecosystem and loss of consensus reality.
See https://arxiv.org/abs/2408.12622 2025-07-14.

^[29] Tucker, J.A. et al., Social Media, Political Polarization, and Political Disinformation: A Review of the Scientific Literature https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3144139 2025-08-02.

^[30]

https://airisk.mit.edu/

. Risk 4, Malicious actors & Misuse, 4.2 Cyberattacks, weapon development or use, and mass harm.

^[31] Davidson, T., Hadshar, R., The Industrial Explosion https://www.forethought.org/research/the-industrial-explosion.pdf 2025-07-14.

^[32] Oesterle, H. „Data Collectors“

,
“Sirion Individual Discussion”

.

^[33] Although this is prohibited under the EU AI Act, it is de facto difficult to control: The Act Texts | EU Artificial Intelligence Act 2025-07-14.

^[34] MacAskill, W., Moorhouse, F., Preparing for the Intelligence Explosion 2506.14863 2025-07-14.

^[35] Spotify has superior know-how within the consortium www.christinebauer.eu, MuRS 2025 2025-08-3.

^[36] E.g. the App juli.co.
The video „Sirion Capital Health“ illustrates the collaboration of people and digital coaches

^[37] Patel, N., https://www.theverge.com/24216748/replika-ceo-eugenia-kuyda-ai-companion-chatbots-dating-friendship-decoder-podcast-interview 2025-06-03

^[38] Bosch Motorcycle stability control (MSC) 2025-07-14.

^[39] Yudkowsky, E., Complex Value Systems are Required to Realize Valuable Futures. https://intelligence.org/files/ComplexValues.pdf. 2035-08-13.

^[40] Yudkowsky, E. Value Theory. High Challenge — LessWrong 2025-08-13.
Oesterle, H., “Measurement of quality of life”

2025-07-14.

^[41] affectiva Affectiva and Emotion AI

https://www.affectiva.com/

2025-08-20.

^[42] Schummeck, J., ad family happiness Familie: Warum Eltern in Deutschland unglücklicher sind als kinderlose Paare | ze.tt 2025-07-14.

^[43] Yudkowsky, E. Initial objective function https://intelligence.org/files/ComplexValues.pdf 2025-08-14.

Bengio, Y., Introducing LawZero - Yoshua Bengio 2025-08-14.
Bengio, Y., Personal and Psychological Dimensions of AI Researchers Confronting AI Catastrophic Risks - Yoshua Bengio 2025-08-14.
UNESCO Editor Missing links in AI governance - UNESCO Digital Library 2025-08-14.

^[44] Joy and suffering are physiological states: https://hubert-oesterle.medium.com/homeostasis-the-engine-for-development-7f879347d3d2 2025-06-30.

^[45] Bertolozzi-Caredio, D. et al., JRC Publications Repository - Risks and vulnerabilities in the EU food supply chain 2025-08-14.

^[46] Hemment, D., Kommers, C., ea. Doing AI Differently: Rethinking the Foundations of AI via the Humanities. White Paper. London: The Alan Turing Institute. 2025-08-18.

^[47] Osterle, H. „Sirion Capital Health“

2025-08-14.
“Sirion Individual Discussion”

2025-08-14.
“Sirion Individual Discussion” https://hubert-oesterle.medium.com/sirion-individual-discussion-3bade1a64452 2025-08-14.
“Sirion AI”

2025-08-14.

^[48] OECD, Bridging the Skills Gap https://www.oecd.org/en/publications/oecd-skills-outlook-2023_27452f29-en.html 2025-08-14.

^[49] MacAskill, W., Moorhouse, F. Industrial explosion https://www.forethought.org/research/preparing-for-the-intelligence-explosion#the-industrial-explosion 2025-08-14.
Bria, F., Timmers, P., Gernone, F. “EuroStack – A European Alternative for Digital Sovereignty” p. 63 https://www.euro-stack.info/docs/EuroStack_2025.pdf 2025-08-19.

^[50] Drexler, E. “AI has unblocked progress toward generative nanotechnologies”

AI has unblocked progress toward generative nanotechnologies

2025-08-12.

^[51] eurostat “Agricultural real factor income per annual work unit (AWU)“ https://ec.europa.eu/eurostat/cache/metadata/fr/sdg_02_20_esmsip2.htm?utm_source=chatgpt.com 2025-08-19.
eurostat Performance of the agricultural sector - Statistics Explained - Eurostat 2025-08-19.

^[52] eurostat https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Businesses_in_the_manufacturing_sector&utm_source=chatgpt.com 2025-08-20.

^[53] Tett, G., Could AI help America out of its debt hole? 2025-08-08.

^[54] UNO A Global Digital Compact — an Open, Free and Secure Digital Future for All | United Nations iLibrary 2025-08-18.

^[55] OECD Steering AI´s Future: Strategies for anticipatory Governance: https://www.oecd.org/content/dam/oecd/en/publications/reports/2025/02/steering-ai-s-future_70e4a856/5480ff0a-en.pdf 2025-08-15.
OECD https://oecd.ai/ai-principles 2025-08-15.
OECD “Recommendation of the Council on Artificial Intelligence“ https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449 2025-08-15.
OECD https://www.oecd.org/content/dam/oecd/en/publications/reports/2025/02/steering-ai-s-future_70e4a856/5480ff0a-en.pdf 2025-08-15.

OECD OECD Legal Instruments 2025-08-15.

^[56] OECD named 30 Organisations of international AI governance systems: https://www.oecd.org/content/dam/oecd/en/about/programmes/strategic-foresight/GSG%20Background%20Note_GSG(2024)1en.pdf/_jcr_content/renditions/original./GSG%20Background%20Note_GSG(2024)1en.pdf 2025-08-19.

^[57] Executive Order 14110 – Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence.

^[58] “Winning the Race. America´s AI Action Plan” https://www.whitehouse.gov/wp-content/uploads/2025/07/Americas-AI-Action-Plan.pdf 2025-08-19.

Field, H., “California is trying to regulate its AI giants – again” California releases long-awaited AI safety report | The Verge 2025-08-19.
Bommasani, R. et al. June-17-2025-–-The-California-Report-on-Frontier-AI-Policy.pdf 2025-08-19.
Hashim, Sh. Tech’s scared of the RAISE Act - by Shakeel Hashim 2025-08-19.

^[59] OECD https://oecd.ai/ai-principles 2025-08-19.

^[60] Barcelo, R., et al. https://www.mwe.com/insights/the-ai-act-the-eus-bid-to-set-the-global-standard-for-ai-regulation/ 2025-06-25.
Schildt, J., et al. Europe must hit pause on the AI Act | Opinion | Sifted 2025-06-25.

^[61] Kokotajlo, D. https://ai-2027.com/research/ai-goals-forecast 2024-08-10

^[62] Järviniemi, O., Hubinger, E. Uncovering Deceptive Tendencies in Language Models: A Simulated Company AI Assistant 2405.01576

^[63] Well-intentioned regulatory initiatives such as those proposed by Digital Humanism failed to achieve global implementation, but have hindered European AI developments: https://dighum.org/open-letter-on-the-urgent-need-to-regulate-digital-technologies/ 2025-05-28.

^[64] MIT AI Risk Repository Risk 6 Socioeconomic & Environmental Harms, 6.1 Power centralization and unfair distribution of benefits

https://airisk.mit.edu/

2025-08-19.

^[65] For some basic ideas see Oesterle, H.
„13 needs“

“Why”

„Values and needs“

“The impact of digital services”

“Needs model”

„Measurement of quality of life”

^[66] Ermut, S. Top 17 Predictions from Experts on AI Job Loss in 2025 2025-08-08.

^[67] See e.g. Christine Bauer (0000-0001-5724-1137) - ORCID 2025-08-14.

^[68] Seydl, J. and Linden, J. use extrapolation to arrive at high, but far lower productivity gains, but do not take into account the scientific progress accelerated by AI How AI can boost productivity and jump start growth | J.P. Morgan Private Bank U.S. 2025-08-02.

^[69] Harari, Y.N. Homo Deus: A Brief History of Tomorrow

^[70] In 2017, the screen time of young people in Germany and the USA already amounted to 3.5 hours per day. For more see:
Fachstelle Öffentliche Bibliotheken NRW Jugend und Medien – Die JIM-Studie 2023 im Überblick – Fachstelle Öffentliche Bibliotheken NRW 2025-08-19.
Livingston, G., How teens spend their time is changing, but boys and girls still differ | Pew Research Center 2025-05-05.

^[71] Kokotajlo does not believe that people will ultimately be able to determine the objective function: https://ai-2027.com/research/ai-goals-forecast 2025-06-14.

^[72] The assumption that a human objective function (a spec in AI 2027 terminology) will be established is optimistic. The OpenAI 2023 alignment team has discussed approaches: Burns, C. et al. https://cdn.openai.com/papers/weak-to-strong-generalization.pdf#page=47.37 2025-06-16.

^[73] Burns, C. et al. https://arxiv.org/pdf/2312.09390 2025-06-14.

^[74] Berkman, J., The Race to Make Designer Babies - The Free Press 2025-08-12.

^[75] The San Francisco Consensus 2025-06-06.

^[76] Oesterle, H., “Homeostasis and Learning”

2025-08-20.

^[77] A distinction between AGI (Artificial General Intelligence) and ASI (Artificial Super Intelligence) according to AI-2027 is not necessary for the social timeline.

^[78] Remaining competition is a very optimistic assumption. There are plausible reasons to believe that there will only be a single superintelligence, see e.g.:
Kokotajlo, D., et al.

https://ai-2027.com/

2025-06-04.

^[79] For the AI market see The EuroStack Initiative pp. 52. https://www.euro-stack.info/docs/EuroStack_2025.pdf 2025-05-30.

^[80] Ngo, R. et al., “ The Alignment Problem from a Deep Learning Perspective” https://arxiv.org/abs/2209.00626 2025-06-02.

^[81] Drexler, E., „AI has unblocked progress toward generative nanotechnologies”

AI has unblocked progress toward generative nanotechnologies

2025-06-14.

^[82] ETFs’ share of global corporate capitalization grows rapidly.
Idzelis, Ch., Why BlackRock sees active ETFs more than quadrupling to $4 trillion by 2030 - MarketWatch 2025-08-19.

Testani, S., “ETFs Defying Gravity: 2024 Flows Surpassed the $1 T

Mark” American Century Investments 2025-07-08.

^[83] Oesterle, H. „13 Bedürfnisse“

2025-08-19.

^[84] As early as 2012, Marshall Brain formulated a utopian, unlimited consumer world alongside a dystopian slave world in his book on living with super-intelligence. However, the needs model of life engineering makes Brain’s vision appear to be a real utopia.

^[85] Oesterle, H.. “Life Engineering. Machine Intelligence and Quality of Life” Springer Nature.

^[86] Davidson, T., et al., “AI-Enabled Coups: How a Small Group Could Use AI to Seize Power” https://www.forethought.org/research/ai-enabled-coups-how-a-small-group-could-use-ai-to-seize-power.pdf 2025-08-12.

^[87] Turner, A. Intrinsic Power-Seeking: AI Might Seek Power for Power’s Sake 2025-07-03
Carlsmith, J., “IsPower-SeekingAIanExistentialRisk?” https://arxiv.org/pdf/2206.13353 2025-07-03.

Center for AI Safety Statement on AI Risk | CAIS 2025-07-03.

^[88] Sam Altman claims the opposite in a possibly business-driven BLOG: The Gentle Singularity - Sam Altman 2025-08-11.

^[89] If you look at the extinction of species in evolution, the expectation for the arrival of this milestone is significantly higher than that of the AI experts surveyed by Grace, K., et al. et al. https://arxiv.org/pdf/2401.02843 S. 15 2025-08-19.

Frequently formulated goals such as immortality, transhumanist abilities, the colonization of alien planets, the uploading of knowledge from humans to humans or from machines to humans, lose their value when humans no longer have a purpose.

^[90] Brynjolfsson, E., Korinek, A., Agrawal, K., A Research Agenda for the Economics of Transformative AI, www.nber.org/papers/w34256 2025-10-18

^[91] The study of forethought.org “Preparing for the Intelligence Explosion” is an excellent collection of opportunities and threats of AI development: https://www.forethought.org/research/preparing-for-the-intelligence-explosion 2025-06-23.
The AI Seoul Summit 2024 summarized the views of delegates from 30 countries in the International Scientific Report on the Safety of Advanced AI:

https://assets.publishing.service.gov.uk/media/6716673b96def6d27a4c9b24/international_scientific_report_on_the_safety_of_advanced_ai_interim_report.pdf 2025-06-17.

^[91] Altman, S., typical argumentation in fundraising Sam Altman 2025-07-26.

^[92] Kokotajlo, D. et al. AI 2027 2025-08-10.

^[93] Acemoglu, D., The Simple Macroeconomics of AI. https://shapingwork.mit.edu/wp-content/uploads/2024/05/Acemoglu_Macroeconomics-of-AI_May-2024.pdf 2025-08-12.

^[94] Kokotajlo, D. et al. AI 2027. 2025-08-10.

^[95] Korinek, A. Scenario Planning for an AGI Future-Anton Korinek 2025-08-12.

^[96] Marcus, G. Taming Silicon Valley: How We Can Ensure That AI Works for Us. MIT Press. Kindle-Version. 2025-08-10.
AI Impacts Wiki Evidence against current methods leading to human level artificial intelligence [AI Impacts Wiki] 2025-08-10.

^[97] Bostrom, N., Superintelligence: Paths, Dangers, Strategies. Oxford University Press 2014.

^[98] Oesterle, H. “Life Engineering Why”

2025-08-10.

^[99] Oesterle, H. „Values and Needs“

2025-08-10.

„Needs Model“

2025-08-10.

^[100] Marshall Brain developed a vivid, perhaps not entirely realistic idea in his book “Manna”.

^[101] Grace, K., et al. pp. 13 https://arxiv.org/pdf/2401.02843 2025-08-10.