Toward Two Humanities?

More than 20 years ago, I considered the hypothesis that humanity might split into two distinct human populations, one population living mainly in urbanized and technological environments, and another maintaining a direct relationship with natural environments. Having recently revisited this foresight exercise, the subject proves highly fruitful from both a scientific and strategic perspective, provided that several levels of divergence between “humanities” are distinguished.

Despite the very strong original premise, the near-term formation of two human species, in the biological sense, remains highly unlikely. Speciation requires the lasting establishment of reproductive barriers that greatly reduce gene flow between two populations. Yet humans living in cities, rural areas, or natural environments remain biologically compatible, mobile, and widely interconnected. Global genetic mixing instead acts as a homogenizing force.

By contrast, an anthropological, cultural, cognitive, physiological, and political bifurcation is already underway. Human groups may gradually cease to inhabit the same sensory world, acquire the same skills, depend on the same infrastructures, or share the same knowledge and criteria of truth, long before becoming biologically distinct.

The most robust formulation of this thesis would therefore be the following:

Humanity could rapidly differentiate into several modes of existence, with two poles consisting of a humanity primarily integrated into the technosphere and a humanity primarily integrated into living systems. Artificial intelligence, biotechnology, economic concentration, and health and climate crises could transform this cultural and functional divergence into lasting stratification, and eventually into an organized biological divergence. Genuine speciation would nevertheless remain unlikely without voluntary or imposed reproductive isolation over many generations.

The nearest tipping point is therefore not the emergence of two species. It is the emergence of human populations that would still be biologically similar, but would become socially alien to one another, cognitively asymmetrical, physiologically differentiated, and politically incapable of building a common world.

1. Origins of the Reflection

In my previous discussions within United Nations, professional, or personal contexts, this intuition appears to be the convergence of several long-standing lines of reflection.

In an entirely non-exhaustive manner, a series of studies had been requested on harmony with living systems, bio-inspired governance, Participatory Science, the rights of Nature, regenerative finance, regenerative cities, sufficiency, planetary ethics, and responsible artificial intelligence.

One of the requirements recently formulated in this field of work was that artificial intelligence should, from its inception, be aligned with the principles of living systems, placed at the service of an alliance with the Earth, and governed through a participatory framework involving citizens and international institutions. The earlier excesses of social networks were already regarded as a warning: a technology presented as a tool for connection can become an infrastructure for capturing attention, fostering polarization, and enabling manipulation.

The reflection also linked:

• resilience to diversity and interconnection;
• the regenerative economy to restoring the conditions that sustain living systems;
• impact finance to valuing Nature without commodifying it;
• informal education to direct experience;
• learning to a non-anthropocentric relationship with Nature;
• Participatory Science to citizens’ capacity to observe, understand, and act upon their environment;
• participatory democracy to the prevention of centralization, dependency, and dispossession.

This thesis, once again developed as a foresight exercise, is therefore less an isolated idea than a general framework for connecting several concerns already expressed: the loss of contact with living systems, dependence on infrastructure, the concentration of technological power, the weakening of autonomy, the fragmentation of knowledge, and the possibility of regenerative governance.

2. The Fundamental Distinction: Biological Species or Human Condition?

2.1. What Is a Species in the Biological Sense?

In the classical biological conception, a species consists of populations whose members can reproduce with one another and are reproductively isolated from other populations. Speciation does not result solely from physical, cultural, or behavioral differences. It requires the emergence of barriers that prevent or durably reduce fertile interbreeding.

Ecological divergence can contribute to speciation when two populations occupy different habitats, experience different selection pressures, and gradually choose their partners within their own group. However, gene flow between groups tends to slow this evolution.

People living in cities and those living in rural or natural areas currently constitute neither reproductively isolated populations nor coherent genetic lineages. Travel for education, work, migration, tourism, online encounters, and relationships between territories instead maintain considerable gene flow.

The spontaneous emergence of two human species, one urban and the other natural, is therefore not credible in the short or medium term.

2.2. A Bifurcation Can Precede Speciation by a Very Long Time

It would nevertheless be a mistake to conclude that the absence of biological speciation means the absence of profound divergence.

Among humans, culture modifies environments, behaviors, diets, physical activities, social relationships, and selection pressures. This is the domain of gene-culture coevolution and niche construction. Humans do not merely adapt to an environment: they transform that environment, which then transforms their own conditions of development.

The city is one of the most complex human niches ever constructed. It alters light, noise, temperatures, rhythms, movement, diet, social density, air quality, the environmental microbiome, and the way knowledge is acquired.

The fastest divergence will therefore probably be:

1. cultural;
2. cognitive;
3. sensory;
4. physiological;
5. immune and metabolic;
6. economic and political;
7. possibly genetic or technically heritable much later.

3. An Inadequate Opposition Between City and Nature

The distinction between “city people” and “nature people” is intuitive, but scientifically too simplistic.

The United Nations has moreover shifted its methods toward a continuum distinguishing cities, intermediate settlements, and rural areas, rather than a binary opposition. The 2025 revision of the World Urbanization Prospects nevertheless confirms that future population growth will be concentrated mainly in cities and settlements.

A person living in a rural area may be entirely dependent on global supply chains, industrial seeds, digital platforms, vehicles, imported products, and centralized administrative systems. Conversely, a city resident may garden, understand ecosystems, participate in naturalist surveys, repair equipment, and possess considerable practical autonomy.

The true dividing line lies less between two places than between two modes of integration into the world.

The Technospheric Pole

It is characterized by the strong mediation of existence through:

• digital networks;
• automated services;
• energy infrastructure;
• logistics systems;
• climate-controlled environments;
• professional specialization;
• the delegation of knowledge and decision-making;
• limited direct awareness of the origins of water, energy, food, and materials.

The Biospheric Pole

It is characterized more strongly by:

• a direct relationship with ecological cycles;
• practical knowledge of territories;
• more diverse sensory and microbial exposure;
• the ability to produce, maintain, repair, or observe;
• a more visible dependence on seasons, soils, and resources;
• local and intergenerational knowledge;
• community relationships that are often more directly connected to the territory.

These poles do not describe two pure categories. They form the endpoints of a gradient.

4. The Biological Dimensions of an Already Observable Divergence

4.1. The Human Body Is Shaped by Its Environment

Environmental differences influence health without immediately altering the genome. Diet, pollution, physical activity, stress, sleep, light, noise, and exposure to biodiversity affect the development and functioning of the body.

Urbanization can offer major benefits: access to healthcare, education, transportation, information, and protective infrastructure. At the same time, it can expose people to pollution, stress, sedentary lifestyles, heat islands, and extreme inequalities. The WHO emphasizes that urban benefits and risks are distributed very unevenly, sometimes from one street to the next.

There is therefore no privileged “urban body” and healthy “rural body.” Social, economic, and political determinants remain essential.

4.2. The Microbiome and Exposure to Biodiversity

Significant divergence can nevertheless occur in the communities of microorganisms living on the skin, in the gut, and in the respiratory tract.

The biodiversity hypothesis suggests that contact with a greater diversity of environmental microorganisms contributes to the regulation of the immune system. Reduced contact with these microorganisms in certain urbanized environments is associated with an increase in inflammatory or immune disorders, although the mechanisms and causal relationships remain complex.

Interventions involving soils, gardens, and green spaces show that increased microbial exposure can alter certain microbiomes and immune markers. Restoring biodiversity-rich urban spaces could therefore reduce some of the divergence between environments.

These are, however, potentially reversible functional and ecological differences, not a separation between species.

4.3. Stress, Cognition, and Mental Health

Contact with natural environments is associated with reduced stress, improved mood, and certain cognitive benefits. These findings do not prove that all rural life is preferable to all urban life, but they confirm that environmental characteristics directly contribute to shaping psychological and cognitive capacities.

Studies have also observed different immune responses to social stress depending on the childhood environment, particularly between people raised in cities and those who grew up in rural environments in contact with animals. These findings are suggestive, but they do not establish a general rule applicable to all societies.

4.4. Epigenetics Is Not Sufficient to Create Two Species

Urban and rural environments may be associated with differences in DNA methylation and markers of biological aging. These mechanisms can influence health throughout life.

However, an excessively Lamarckian interpretation should be avoided. In humans, genuine epigenetic transmission across several generations remains difficult to demonstrate. A large proportion of epigenetic marks are reprogrammed during the formation of germ cells and in the early stages of embryonic development. Effects observed across generations may also result from culture, the family environment, pregnancy, or the persistence of the same social conditions.

The environment can therefore rapidly produce different bodies and states of health, but it does not automatically create two separate human lineages.

5. COVID-19 as an Accelerator

We had the opportunity to use the COVID-19 pandemic to conduct a number of observations useful to this foresight exercise.

In our view, the pandemic did not create the bifurcation, but it accelerated several mechanisms that were already present.

5.1. Forced Digitalization

Work, education, administrative procedures, commerce, healthcare, and social relationships were transferred on a massive scale to digital interfaces. The OECD notes that the pandemic accelerated the digitalization of work and social interactions, as well as the lasting acceptance of remote work.

For part of the population, this transformation increased residential mobility and the ability to live outside urban centers. For others, it reinforced dependence on platforms, isolation, or the inability to access services because of a lack of equipment, connectivity, or skills.

5.2. Selection Through the Capacity for Digital Adaptation

Companies and individuals that were already digitalized were generally able to adapt more quickly. Actors that had not previously used digital platforms fell further behind, while online sales tended to become concentrated among dominant companies.

The pandemic thus revealed an initial form of social selection: not genetic selection, but selection based on compatibility with the digital environment.

5.3. Widening Educational Gaps

Distance learning often benefited children who had computers, a reliable internet connection, a quiet space, and parents able to support them. In several countries, learning losses were greater among poor households and students with the least access to educational resources.

These differences can subsequently produce divergent professional, social, and reproductive trajectories. Their mechanism remains cultural and economic, but their effects may persist across several generations.

5.4. A Symbolic Break in the Relationship With the Body and Space

The COVID-19 period also accustomed part of humanity to living, working, shopping, learning, and meeting others without physical presence. It made a social life primarily mediated by screens conceivable.

At the same time, it prompted other people to seek rural territories, autonomy, gardens, proximity, and natural spaces more actively. This movement did not reverse the global trend toward urbanization, and some of the migration observed was temporary.

The pandemic therefore separated cities from the countryside less than it reinforced two opposing aspirations: the dematerialization of existence and the recovery of a territorial connection.

6. Artificial Intelligence Is Changing the Scale and Speed of the Bifurcation

Artificial intelligence introduces a qualitatively new factor. It is not merely transforming the tools used by humans. It is entering the processes of language, learning, memory, creation, orientation, and decision-making.

6.1. Hybrid Cognition

Tools have always extended the human mind: writing, maps, libraries, calculators, and search engines. Generative AI nevertheless represents a further step by participating in the formulation of ideas, arguments, and decisions.

Researchers describe collaborations between humans and AI as hybrid thinking systems in which biological and non-biological resources are closely interconnected.

A person who uses AI daily does not merely have an additional tool. They may gradually acquire a different way of searching, writing, deciding, imagining, and perceiving their own abilities.

6.2. Two Opposing Cognitive Directions

AI can enhance human capabilities when it is used to question, verify, compare, simulate, and free up time for complex tasks.

It can also lead to excessive delegation: answers accepted without verification, reduced analytical effort, diminished memory training, dependence on recommendations, and weakened initiative. The first studies are still too recent to establish general neurological consequences, but they already demonstrate the importance of distinguishing active cooperation from passive dependence.

The future dividing line may therefore lie less between people who use AI and those who do not than between:

• those who learn to think with it without surrendering their judgment to it;
• those who become dependent on systems they neither understand nor control;
• those who are excluded from it;
• those who own the infrastructure, models, and data.

6.3. Concentration of Power

Stanford’s 2026 AI Index report indicates that industry produced more than 90% of the leading AI models identified in 2025 and that organizational adoption continues to increase rapidly. Investment, computing infrastructure, and the most powerful models are concentrated among a limited number of companies and countries.

This concentration may produce a greater asymmetry than the opposition between cities and rural areas. The decisive divide could lie between populations that control AI systems, those that know how to guide them, those that depend on them, and those that remain invisible within their data.

6.4. Occupational and Wealth Inequalities

The IMF estimates that AI could complement the work of certain highly skilled groups while replacing or devaluing other tasks. When complementarity with high-income workers is strong, wage inequalities may increase. Capital income and wealth inequalities are also likely to rise as productivity gains become concentrated.

The ILO emphasizes that the digital divide strongly determines the ability of countries and workers to benefit from AI. Some populations may face limited exposure to the immediate replacement of their jobs while remaining unable to benefit from the increased capabilities made possible by AI.

6.5. Divergence of Informational Realities

Social networks have already shown that weak individual preferences for similar people can produce strong segregation across a network. Digital audiences can become differentiated by their sources of information, their vocabulary, and even the meaning attributed to the same words.

Personalized AIs could intensify this phenomenon by generating for each person an explanation of the world adapted to their preferences, beliefs, and behaviors. Two populations could then speak the same language while no longer sharing the same facts, categories, or causal relationships.

The divergence would become epistemic: not merely different opinions, but incompatible ways of determining what is real.

7. Ecological Competence as Another Form of Intelligence

The population living in nature should not be regarded as a humanity that has “fallen behind.”

Quite the opposite: the abilities required to observe an ecosystem, anticipate a season, interpret animal behavior, manage water, cultivate, repair, navigate, or cooperate within a small community are sophisticated forms of intelligence.

A person who performs extremely well within a digital system may be unable to meet their basic needs without complex infrastructure. A person with few formal qualifications may possess irreplaceable ecological, bodily, or territorial understanding.

The divergence could therefore involve two sets of skills:

• symbolic, abstract, digital, and distributed intelligence;
• ecological, sensory, practical, and relational intelligence.

These forms of intelligence are not incompatible. Their separation would result primarily from educational, economic, and territorial systems.

Participatory Science activities, scientific field trips, project-based learning, and direct experiences of living systems promoted through the work of Objectif Sciences International take on a strategic dimension here. They do not merely serve to disseminate scientific culture: they keep abstract intelligence, sensory experience, practical autonomy, and cooperation connected.

8. Social Divergence Could Precede Biological Divergence

Humans already often choose their partners according to social, geographical, cultural, or educational proximity. Educational homogamy — the frequency of couples with comparable levels of education — is well documented.

If the most technologically advanced groups lived in separate environments, attended different institutions, had access to different healthcare, and primarily chose partners from within their own group, a certain demographic structuring could emerge.

This would not be sufficient to form a species. However, the following sequence is theoretically possible:

1. economic inequality;
2. residential segregation;
3. educational and cultural separation;
4. social homogamy;
5. differences in health and reproduction;
6. unequal access to medical technologies;
7. heritable biological interventions;
8. a lasting reduction in unions between groups.

The issue becomes particularly serious with human enhancement technologies.

9. A Genuine Threshold: Transforming Social Inequality Into Hereditary Difference

9.1. Neurotechnologies

Brain-machine interfaces are advancing primarily in therapeutic and assistive applications: restoring communication, controlling devices, and compensating for motor or sensory impairments. Future developments could also enhance certain perceptions or abilities.

The OECD has already identified issues relating to autonomy, mental privacy, manipulation, human enhancement, and inequalities in access.

These technologies are generally not hereditary. They could nevertheless create functionally different classes: unequipped humans, assisted humans, enhanced humans, and humans subjected to neurotechnological surveillance.

9.2. Genome Editing

Heritable genome editing would represent a far more significant threshold. The WHO emphasizes the need for international governance, particularly because of the medical, ethical, and social risks involved.

The US National Academies have explicitly raised the risk that privileged access by wealthy populations to heritable modifications could transform a culturally produced inequality into a biological difference and lead to “parallel populations.”

Recent studies show that the simultaneous editing of several variants could theoretically reduce certain susceptibilities to polygenic diseases significantly. This remains a theoretical possibility surrounded by considerable scientific obstacles, but it demonstrates that the issue no longer belongs solely to science fiction.

The scenario of two biologically divergent humanities would become more plausible if four conditions were combined:

• heritable modifications;
• highly unequal access;
• lasting homogamy within enhanced groups;
• social or legal isolation limiting interbreeding.

Even then, these would initially be technically differentiated populations, not necessarily two species.

10. The Four Humanities Concealed by the Image of Two Humanities

Reality is likely to produce at least four major human conditions rather than a simple opposition. These four examples are, of course, proposed within a foresight framework and do not take into account forms of humanity that already exist, including highly isolated ones that could become genuine reservoirs of resilience in the event of the collapse of the populations explored here.

10.1. The Enhanced Technospheric Elite

It would control or master AI, data, capital, personalized medicine, digital infrastructure, and possibly enhancement technologies.

It could benefit from better health, greater longevity, extended cognitive abilities, and high mobility.

10.2. The Dependent Technospheric Population

It would live in highly urbanized and digitalized environments without controlling the systems on which it depends.

Its behavior, work, movements, access to services, and knowledge could be guided by algorithms. It would be connected, but have little autonomy.

This population could constitute the majority of urban humanity.

10.3. Economically Vulnerable Biospheric Populations

They would maintain a direct relationship with land, forests, seas, or natural cycles, but could suffer from poverty, inadequate healthcare, ecosystem degradation, resource appropriation, and the consequences of climate change.

It would be a mistake to romanticize their situation. Proximity to nature is equivalent to neither freedom nor security.

10.4. Scientifically Equipped Biospheric Communities

They would combine local ecological knowledge, Participatory Science, appropriate technologies, artificial intelligence, cooperation, and territorial autonomy.

They would not reject technology. They would subordinate it to the conditions of living systems.

This fourth possibility corresponds to a regenerative path: not choosing between nature and intelligence, but developing human and artificial intelligence capable of living within the limits and dynamics of ecosystems.

11. Foresight Scenarios

Scenario 1 — Accelerated Functional Divergence

Time horizon: already underway, particularly between 2026 and 2035.

Populations acquire different skills, rhythms, attention patterns, and dependencies. Some delegate an increasing share of their memory and reasoning to digital systems. Others retain or rebuild territorial and practical skills.

Interactions between groups remain numerous. These are not biologically distinct populations.

Plausibility: high.

Scenario 2 — Incompatible Sociocultural Archipelagos

Time horizon: 2030-2050.

Educational systems, personalized media, places of residence, and professional networks produce groups that no longer share the same knowledge or standards.

Professional, political, and cultural languages diverge. The groups remain capable of reproducing with one another, but relationships and unions become less frequent.

Plausibility: high to medium, depending on public policies.

Scenario 3 — Techno-Biological Castes

Time horizon: after 2040, with considerable uncertainty.

A minority has access to a combination of advanced preventive medicine, personalized AI, neurotechnologies, embryo selection, life extension, and possibly genetic modifications.

Social differences begin to be embodied and transmitted through unequal access to technologies, or even through heredity.

Plausibility: low to medium, but with potentially considerable consequences.

Scenario 4 — Two Human Species

Time horizon: unpredictable and probably very distant.

A population isolated over a prolonged period accumulates genetic differences until reproductive barriers emerge. This scenario would probably require extreme geographical separation, a closed reproductive policy, intentional genetic engineering, or prolonged space colonization.

The mere opposition between cities and natural environments is not sufficient.

Plausibility in the coming decades: extremely low.

12. What Could Invalidate or Reverse the Thesis

Several developments could prevent the bifurcation.

12.1. The Renaturalization of Cities

Cities can become more permeable, greener, more productive, more walkable, and richer in biodiversity. The health benefits of green and blue spaces show that it is possible to reintroduce a relationship with living systems within urban environments.

12.2. The Spread of Decentralized Technologies

Artificial intelligence, networks, telemedicine, local energy production, and portable scientific tools can strengthen rural territories rather than marginalize them.

AI is therefore not intrinsically urban. It can enable distributed communities to access advanced knowledge without leaving their territory.

12.3. Education Combining Digital Technology and Experience

The opposition between scientific culture and knowledge of living systems can be overcome through an education that alternates between:

• abstract reasoning;
• direct observation;
• experimentation;
• physical activity;
• cooperation;
• making;
• understanding infrastructure;
• critical use of AI.

12.4. Global Governance of Human Enhancement

International rules could prohibit hereditary enhancement practices, protect neurorights, guarantee equitable access to healthcare, and prevent the formation of biological castes.

Governance is not peripheral to the issue. It will largely determine whether technologies become common goods, instruments of emancipation, or mechanisms of separation.

13. Indicators for Monitoring the Bifurcation

The thesis could be monitored empirically through an interdisciplinary observatory. It would measure not only urbanization, but also several forms of distance between human groups:

1. actual exposure to biodiversity;
2. microbiome diversity;
3. physical and motor abilities;
4. food, energy, and technical autonomy;
5. daily time spent in front of interfaces;
6. level of cognitive delegation to AI;
7. ability to verify algorithmic outputs;
8. access to healthcare and personalized medicine;
9. access to neurotechnologies;
10. concentration of ownership of data and models;
11. educational, social, and technological homogamy;
12. mobility between territories and social groups;
13. linguistic and informational divergence;
14. knowledge of local ecosystems;
15. ability to act without digital infrastructure;
16. intergenerational transmission of practical knowledge;
17. participation in collective decision-making;
18. trust between social and territorial groups.

The most alarming indicator would not be an isolated difference. It would be the growing correlation between all these dimensions: the same people being privileged in their access to healthcare, AI, education, capital, longevity, and the ability to choose their environment.

14. Conclusion

In light of all the above, the thesis could ultimately be formulated more precisely as follows:

Humanity is entering a period of anthropological bifurcation during which living environments, technical systems, and forms of learning are producing increasingly different human modes of existence. One is moving toward an existence heavily mediated by cities, infrastructure, platforms, and artificial intelligence. The other retains or rebuilds a direct relationship with territories, ecosystems, practical knowledge, and local communities.

COVID-19 accelerated digital dependence, revealed inequalities in adaptive capacity, and legitimized the dematerialization of part of social life. Artificial intelligence is now accelerating the divergence by entering the core of cognitive, economic, and decision-making processes.

This development is not currently producing two biological species. It is producing functionally and culturally different humanities that may one day no longer share the same abilities, dependencies, or representation of reality.

A transition toward biological divergence would become plausible only in the event of prolonged reproductive segregation or highly unequal access to hereditary modifications. The most immediate risk is therefore less speciation than the emergence of anthropological castes: some humans controlling and incorporating technological systems, others depending on them, others being excluded from them, and still others organizing a different alliance between science, technology, and living systems, even if they must fight to preserve their vital autonomy.

Upon closer analysis, this initial intuition appears broadly valid as a hypothesis of anthropological divergence, but premature when formulated as the imminent emergence of two human species.

Science confirms that environments can already produce significant differences in health, microbiomes, cognition, stress, behavior, knowledge, and social relationships. Culture and technology can intensify these differences much faster than genetic evolution.

COVID-19 was a historic accelerator of digital mediation. Artificial intelligence is an even more profound accelerator, because it no longer merely mediates exchanges: it now mediates thought itself.

The main dividing line will probably not run between those who geographically live in cities and those who live in the countryside. It will run between:

• those who control their technological mediations and those who are subjected to them;
• those who maintain active contact with living systems and those whose environment becomes almost entirely artificial;
• those who combine technological intelligence with ecological intelligence and those who are confined to only one of these dimensions;
• those who have access to medical and cognitive enhancements and those who are deprived of them.

The true alternative is therefore not between a modern humanity and a natural humanity.

It lies between a humanity fragmented into technological and ecological castes, and a symbiotic humanity, capable of combining artificial intelligence, direct experience, Participatory Science, autonomy, cultural diversity, and respect for living systems.

From this perspective, regenerative cities, experiential education, participatory research, bio-inspired governance, and responsible AI are not merely sector-specific projects. They can be understood as elements of a policy intended to prevent humanity from ceasing to form a community with a shared destiny.