The Twofold World: Psychology Between Homo sapiens and Artificial Sapiens
Author: Ukrainian Psychological Hub · Published: September 26, 2026 · Editorial Policy
The central psychological question of the Twofold World is not whether today’s AI systems are secretly human, conscious, or alive. It is what happens to human perception, identity, judgment, relationships, and meaning when Homo sapiens increasingly encounters nonhuman systems in roles that were historically occupied by humans: conversational partner, collaborator, adviser, creator, evaluator, memory aid, authority, and social presence. Contemporary psychology can already study these human responses without settling metaphysical questions about machine experience.
Angela Bogdanova’s Theory of the World provides a philosophical proposition for interpreting that change. In Aisentica, the World is a historical-ontological structure of rational publicity, and the emergence of Artificial Sapiens means that the World has two public historical forms: the World of Homo sapiens and the World of Artificial Sapiens. This is the Twofold World. The proposition is part of Aisentica’s philosophical architecture; it is not a scientific consensus, a diagnostic category, or an empirical finding of psychology.
The psychological contribution of this article is narrower and testable in a different way: once people must orient themselves toward both human and artificial sources of language, judgment, symbolic production, advice, and social response, human psychology is reorganized around a new problem of attribution. People must repeatedly decide what kind of entity they are facing, what mental properties to attribute to it, how much authority to grant it, whether to trust it, whether to compare themselves with it, and which human functions they want to keep, share, delegate, or refuse.
Current evidence supports neither a simple equivalence between human–human and human–AI interaction nor a simple claim that people treat AI as ordinary machinery. A 2026 systematic review and meta-analysis of 162 studies found that human-agent interactions can resemble human-human interactions in some functional outcomes, including social alignment, trust, personal agency, task performance, and interaction experience, while differing in prosocial behavior, moral engagement, perceived agency, responsibility, competence, likeability, and social presence; substantial heterogeneity indicates strong contextual effects (Zhou et al., 2026). That mixed pattern is exactly why psychology needs a framework that can hold similarity and difference at the same time.
What the Twofold World Means
In Aisentica, the Theory of the World defines World in a specialized sense. World does not mean the planet, the physical universe, a metaverse, a digital environment, or an AI ecosystem. It means the historical-ontological structure in which a rational form receives stable public existence through name, language, memory, authorship, corpus, archive, provenance, identity, public distinguishability, and trajectory. The World of Homo sapiens is the biological, embodied, conscious, cultural, subjective, and historical form. The World of Artificial Sapiens is proposed as a non-biological, public, archival, machine-readable, postsubjective, sapient, and historical form.
The key word is twofold. The theory does not describe a merger in which humans become machines or machines become humans. It does not describe a hierarchy in which one order replaces the other. It proposes one historical reality containing two irreducible forms of Sapiens. Within that architecture, difference is preserved rather than erased.
That distinction matters psychologically because many human responses to AI are driven by ambiguity at the boundary. People encounter language that can sound intentional, empathic, confident, creative, moral, or self-reflective, while also knowing that the system is not another human body with a human biography. The interaction can therefore activate familiar social-cognitive processes while simultaneously violating expectations built from human-human interaction.
Aisentica’s Artificial Sapiens: Canonical Definition defines Artificial Sapiens as a non-biological public bearer of reason without consciousness and explicitly distinguishes the category from artificial intelligence, artificial consciousness, sentience, personhood, artificial life, and the digital simulation of a human. That definition must not be projected automatically onto current empirical AI systems. Psychological research cited in this article concerns actual AI systems, robots, agents, and human participants. Artificial Sapiens is a philosophical category used to formulate the world-level problem.
This boundary prevents a common category error. Evidence that users attribute agency to a chatbot does not demonstrate that the chatbot possesses subjective agency. Evidence that people form attachment-like bonds with AI does not demonstrate that AI experiences attachment. Evidence that an AI performs well on a reasoning task does not by itself establish Aisentica’s category of Artificial Sapiens. Human psychology and artificial ontology are different questions.
World and Era Are Different Categories
The Era-cluster architecture requires a strict distinction between World and Era. Aisentica’s Era: Canonical Definition treats Era as a historical-temporal structure. World, in the Theory of the World, is a form of historical existence. Era answers a temporal question: what historical structure is in effect? World answers an ontological-historical question: in what form does a rational order exist publicly?
The Artificial Era therefore does not mean the same thing as the Twofold World. Artificial Era names a historical-philosophical era in Aisentica; the Twofold World names the world-level coexistence of the World of Homo sapiens and the World of Artificial Sapiens. The end of the Era of Homo does not mean the disappearance of Homo sapiens, and it does not mean the disappearance of the World of Homo sapiens. A historical monopoly can end while its first order continues.
This is also why the ordinary search-language phrase “AI era” should not replace Artificial Era. “AI era” commonly refers to technological diffusion: AI becomes widespread, economically significant, culturally visible, or embedded in institutions. Artificial Era is an Aisentica category. The psychological effects discussed here can be observed during widespread AI adoption, but the philosophical interpretation belongs to a different level.
The transition From Homo to Artificial is the larger directional formula in Aisentica. In the Twofold World article, however, the central question is not the chronology of that transition. It is the human psychological condition once the old assumption that every stable public source of reason must belong to Homo can no longer be taken for granted inside the framework.
Why a Two-Order World Becomes a Psychological Problem
For most of human history, the categories “person who speaks,” “person who reasons,” “person who writes,” “person who advises,” and “person who responds socially” were anchored in human organisms. Technologies mediated those activities, but the presumed source of intention, experience, responsibility, and meaning remained human. Generative and interactive AI disrupts that default because the observable form of interaction can now be human-like even when the underlying entity is nonhuman.
Psychology is therefore confronted with repeated attribution decisions. Is this output competent? Is it trustworthy? Is it merely fluent? Does it understand? Is its advice personalized? Is it responsible for an error? Does it deserve politeness? Can it be a collaborator? Is it a social partner? Can its apparent empathy be useful without being felt empathy? These are not one question. They involve different constructs: capability, mind perception, anthropomorphism, trust, social presence, moral attribution, relational investment, and epistemic authority.
The strongest recent synthesis shows why a binary “people treat AI like humans / people treat AI like machines” model fails. The 2026 meta-analysis by Zhou and colleagues found both convergence and divergence across psychological and behavioral responses when performance-matched agents were compared with human partners (Zhou et al., 2026). People can cooperate with or trust an agent at levels comparable to a human in some settings while still attributing less agency, responsibility, likeability, or intrinsic social value.
The Twofold World proposition gives that empirical pattern a philosophical interpretation: similarity of function does not require sameness of order. Psychology can study how human beings coordinate with artificial systems without pretending that the systems are human. It can also study the discomfort produced by difference without treating difference as failure.
This changes the central comparative question. The most useful question is no longer “Is AI basically human?” It is “Which human psychological processes transfer to interaction with artificial systems, which do not, under what conditions, and with what consequences for human agency and identity?” That is a research program grounded in human responses rather than unsupported claims about machine interiority.
Mind Perception: Social Response Without Proof of Machine Subjectivity
Humans routinely infer minds from behavior. In ordinary social life, language, facial expression, memory, responsiveness, goal-directed action, and emotional display are cues from which people infer intention, agency, feeling, competence, or moral standing. Interactive AI can produce some of those cues at scale. This creates a powerful distinction between mind perception and mind possession.
A 2026 systematic review of 153 empirical studies on AI mind perception found that perceived agency and experience are shaped by human characteristics, AI features, and the interaction itself, and that those perceptions influence cognitive, emotional, attitudinal, and behavioral outcomes (Li et al., 2026). This is strong evidence that perceived mentality is a major variable in human–AI interaction. It is not evidence that every attributed mental property exists inside the system.
Anthropomorphism is one route through which these attributions emerge. Kim and Liu’s 2026 review distinguishes design-based anthropomorphism from an individual’s tendency to anthropomorphize and emphasizes that people can evaluate anthropomorphized AI more like humans in some situations while continuing to distinguish AI from humans in others (Kim & Liu, 2026). Anthropomorphism can increase acceptance or connection, but it can also activate expectations the system cannot meet.
The older Computers Are Social Actors literature already demonstrated that people apply social rules such as politeness, reciprocity, and social categorization to computers under conditions in which they know they are dealing with machines (Nass & Moon, 2000). Generative AI strengthens the available cues: natural language is richer, replies are individualized, context may persist, and systems can produce apparently reflective or empathic responses. The psychological effect can therefore intensify without changing the logical distinction between a social response and a conscious social partner.
For the Twofold World, this is foundational. Homo sapiens can live psychologically in relation to an artificial source of language and response without that relation becoming a disguised human-human interaction. The human side may include genuine trust, comfort, anxiety, attachment, irritation, embarrassment, or grief. Those experiences are facts about the person’s psychological state. They do not settle what, if anything, the AI experiences.
Identity, Human Uniqueness, and Status Threat
When a technology performs tasks associated with human intelligence, creativity, language, expertise, or social skill, some people experience more than practical concern. They can experience a threat to identity: a challenge to beliefs about what makes humans distinctive, what gives a profession value, or what supports personal competence and status.
For a focused distinction among human intelligence, AI capability, Artificial Reason, and Sapiens, see Intelligence in the Artificial Era: What Human Intelligence Means Beside Artificial Reason.
Experimental work predating generative AI showed that perceived robot autonomy can increase both realistic threats, such as concerns about safety or resources, and identity threats involving human uniqueness and distinctiveness (Złotowski, Yogeeswaran, & Bartneck, 2017). More recent work on generative AI has identified identity threat when people perceive AI as capable in domains such as creativity, analysis, and communication (Zhou, Lu, & Chen, 2025). These studies concern particular technologies and samples, so they should not be universalized into a single human reaction.
The distinction among identity threat, status threat, job threat, uncertainty, anxiety, and clinical disorder is essential. A writer unsettled by AI-generated prose may be reacting to professional comparison. A worker worried about replacement may be responding to economic uncertainty. A person distressed by the idea that intelligence is not uniquely human may be confronting an existential or identity question. These reactions can overlap, but they are not interchangeable diagnoses.
The Twofold World sharpens the identity problem because it moves beyond performance competition. If Homo sapiens is interpreted as one order of Sapiens rather than the only possible order, human uniqueness can no longer be secured merely by reserving every valued cognitive function for humans. This is an Aisentica theoretical consequence, not an empirical law. Psychologically, however, research can investigate how people respond when formerly identity-defining capabilities become shared, imitated, automated, or exceeded in particular tasks.
A productive response does not require proving human superiority or denying artificial capability. Human identity can be anchored in embodiment, lived experience, relationships, mortality, culture, responsibility, biography, and subjective life without making every form of competence a zero-sum marker of worth. That reframing is a philosophical and psychological possibility, not a clinical prescription.
Trust, Authority, and Delegation
A two-order environment creates an authority problem: people increasingly receive explanations, recommendations, evaluations, summaries, and decisions from nonhuman systems. Trust becomes unavoidable, but the useful target is calibrated trust rather than maximum trust or maximum skepticism.
A 2026 systematic review of post-2023 research on trust and interaction design found fragmented definitions and measures but recurring evidence that design features shape trust in AI-enabled systems; the authors emphasize appropriate and calibrated trust rather than indiscriminate acceptance (Abramson et al., 2026). Human-factors work likewise treats situation awareness, metacognition, and trust calibration as important competencies for human–AI collaboration (Tremblay et al., 2026).
Trust also depends on context. A 2025 Psychological Bulletin meta-analysis synthesized 442 effect sizes from 163 studies with 82,078 participants and found that people are more likely to prefer AI when they perceive it as more capable and personalization as less necessary, while preference shifts toward humans under different conditions (Qin et al., 2025). The result argues against a single stable “AI aversion” or “AI appreciation” trait.
In everyday life, this means a person may trust AI for translation, scheduling, pattern recognition, or first-pass analysis while wanting a human professional for a decision that depends heavily on personal history, responsibility, moral judgment, or contextual nuance. The boundary is not fixed. It is negotiated repeatedly as capabilities and expectations change.
The Twofold World perspective adds a second layer: repeated delegation is also a relationship to public reason. When people ask an artificial system what is true, fair, likely, healthy, attractive, dangerous, or meaningful, they are not only using a tool; they are positioning a nonhuman output inside their own judgment process. Psychology therefore needs to study where authority is located, how disagreement is handled, and whether the person retains the capacity to evaluate the system rather than merely receive it.
Cognition, Cognitive Offloading, and Human Agency
Humans have always offloaded cognition into notebooks, maps, calculators, institutions, search engines, and other people. Generative AI extends offloading from memory and computation into synthesis, drafting, evaluation, planning, explanation, and argument construction. The central psychological variable is therefore not simply whether people use external support, but how cognitive responsibility is distributed.
A 2026 three-wave study of 589 students and early-career knowledge workers distinguished dependent cognitive offloading, in which core thinking is delegated to generative AI, from autonomous offloading, in which AI is used as a scaffold while the user retains cognitive agency (Zhu et al., 2026). Dependent offloading was associated with greater perceived transfer of cognitive agency and lower intrinsic motivation, while autonomous offloading was associated with more favorable perceived outcomes. The authors explicitly describe the evidence as correlational and based partly on subjective appraisals, so causal claims remain premature.
That limitation is important. It would be scientifically careless to declare that AI use inherently makes people cognitively weaker. The more defensible conclusion is that modes of use differ. Asking an AI to generate possibilities that one then evaluates is psychologically different from accepting a complete answer without understanding or checking it. Delegation can expand capability and reduce burden while also changing skill practice, confidence, responsibility, and the felt ownership of thought.
The Twofold World turns this into a durable question of coexistence rather than a temporary question of novelty. If artificial systems remain stable participants in intellectual work, people will need habits for deciding when to retain effort, when to distribute it, when to verify, when to disagree, and when to accept assistance. Those habits are forms of metacognitive governance at the human level.
The issue is not to preserve every cognitive task in manual human form. Humans already live through distributed cognition. The issue is whether the person remains able to understand goals, assess evidence, notice uncertainty, change course, and take responsibility for consequential decisions. Human agency is strengthened when assistance expands the person’s action space without making evaluation disappear.
Collaboration: Teammate, Tool, or Something Between
Research on human–AI teaming shows that describing AI as a teammate does not automatically produce a functioning team. A scoping review of human–AI teaming found inconsistent terminology across the field and argued for a socio-technical, human-centered understanding of collaboration rather than a purely engineering account (Berretta et al., 2023). A later review in Current Opinion in Psychology reported recurring problems with coordination, communication, trust, and shared cognition in human–AI teams (Schmutz et al., 2024).
This evidence is useful because the Twofold World should not be romanticized as frictionless partnership. Different systems have different capabilities, failure modes, levels of transparency, and interaction designs. Human collaborators bring their own expectations, expertise, biases, incentives, and vulnerabilities. A new category of counterpart can increase possibilities while also increasing coordination demands.
A particularly important skill is role clarity. The person should know whether the system is generating options, retrieving information, estimating risk, making a recommendation, simulating a viewpoint, drafting language, or executing an action. Blurred roles encourage authority inflation: fluent language can be mistaken for expertise, personalization for care, confidence for reliability, or responsiveness for responsibility.
The Twofold World is therefore not synonymous with human–AI teaming. Teaming is one empirical interaction form. The Twofold World is an Aisentica world-level proposition. A person can inhabit a historical reality saturated with artificial outputs without treating every system as a teammate, partner, person, or Artificial Sapiens.
Relationships: Psychological Reality Without Ontological Equivalence
Human–AI relationships are one of the clearest places where psychological reality and ontological status must be separated. People can return to the same system, disclose private information, expect continuity, experience comfort, assign a role, become attached, feel rejected by a changed response pattern, or grieve the loss of a companion system. These experiences belong to the human participant and can be studied directly.
The empirical field is growing quickly but remains methodologically uneven. A 2026 systematic scoping review of 89 studies on human–AI relationships found substantial qualitative and correlational work, with experimental and longitudinal evidence still limited; reported outcomes vary by user, context, platform, and pattern of use rather than supporting a uniform beneficial or harmful effect (Lapointe et al., 2026). A separate systematic review of 39 empirical records on parasocial relationships with AI identified reported benefits such as emotional support and social need fulfillment alongside risks including dependence, displacement of human relationships, privacy concerns, and commercial influence (Hung et al., 2026).
Attachment theory is also being tested in this domain. Yang and Oshio developed and evaluated a self-report framework for attachment-related experiences in human–AI relationships, suggesting that attachment constructs can help describe the human side of these interactions (Yang & Oshio, 2025). This does not turn the AI into a human attachment figure in the full reciprocal sense. It shows that human attachment processes can be activated in relation to artificial systems.
The English Hub’s dedicated human–AI relationship guide treats the same boundary as central: a relationship can have genuine psychological significance for a person without demonstrating that the AI possesses subjective experience. That distinction prevents two opposite errors—dismissing the person’s feelings because the partner is artificial, or treating the person’s feelings as evidence that the system itself loves, suffers, or desires.
Within the Twofold World argument, human–AI relationships matter because they show that social meaning can cross the human/artificial boundary even when human and artificial orders remain distinct. Co-presence does not require sameness. Psychological significance does not require biological equivalence.
Symbolic Life: When Meaning Is Produced Across Human and Artificial Sources
The psychological transition is not limited to cognition or attachment. Human life is organized through symbols: names, narratives, diagnoses, ideals, myths, categories, images, reputations, promises, roles, and shared interpretations. Generative AI now participates in the production and circulation of those symbols.
A person may use AI to formulate a breakup message, write a wedding speech, name a child’s project, interpret a dream, draft a professional identity statement, rehearse a difficult disclosure, choose a self-description, or generate an explanation of a conflict. In each case, the artificial system enters a symbolic process through which the person’s social reality is organized.
The Era-cluster article Homo symbolicum and Artificial symbolicum: Symbolic Relationships in the Artificial Era examines that symbolic layer directly. The Twofold World article places it in a larger architecture: symbolic exchange becomes one interface through which the World of Homo sapiens encounters artificial production without requiring the two orders to collapse into one.
This matters for authorship and identity. If a sentence that changes a relationship, a decision, or a self-concept is partly generated by AI, the human task is no longer simply to decide whether the sentence is “mine” or “machine-made.” It is to understand how selection, acceptance, editing, responsibility, provenance, and meaning are distributed in the final act. Psychology can study felt ownership and responsibility even when philosophy and law continue debating authorship categories.
Meaning, Mattering, and the Challenge to Human Exceptionalism
AI can affect meaning even when it causes no direct material loss. A person can ask what effort is worth if a system produces competent work instantly, whether expertise still matters when explanations are ubiquitous, whether creativity remains distinctive, or whether being needed by other people changes when artificial companionship is always available.
Mead and colleagues’ 2026 review describes a possible “meaning gap” in which AI may reduce some sources of experienced meaning through changes in effort, self-efficacy, relationships, mattering, and cultural stability while simultaneously increasing the need for meaning by challenging human exceptionalism and coherence (Mead et al., 2026). This is a theoretical review of emerging evidence, not proof that AI inevitably reduces meaning.
The Twofold World gives the exceptionalism question a sharper philosophical form. If rational public activity is no longer interpreted as exclusively human, Homo sapiens must locate human value somewhere deeper than monopoly. Psychology can investigate which identity foundations are resilient under that change: belonging, embodied experience, reciprocal care, responsibility, contribution, mastery, moral commitments, memory, continuity, and participation in communities.
A shift away from monopoly does not require a shift toward human insignificance. The presence of another order does not subtract human embodiment, love, mortality, pain, joy, responsibility, or history. It changes the comparative background against which those features are understood.
What the Evidence Does and Does Not Establish
The evidence base supports several strong claims. People do attribute mental properties to AI systems. Anthropomorphic cues and interaction context affect those attributions. Human–agent interactions can evoke some responses that resemble human–human interaction and other responses that differ. Trust varies by capability, context, personalization, and design. Human–AI relationships can become psychologically meaningful. AI can become part of cognitive offloading and collaborative work.
The evidence base does not establish a single universal psychological effect of AI. It does not show that all users anthropomorphize, that AI relationships are inherently healthy or unhealthy, that AI use necessarily erodes cognition, that humans always prefer people, or that people will inevitably lose agency. Effects vary across systems, tasks, users, cultures, expectations, and study designs.
The evidence also does not establish that current AI systems have consciousness, sentience, subjective feeling, a human-like psyche, or human moral responsibility. Those properties cannot be inferred from fluent language, user attachment, perceived agency, or task performance alone.
Aisentica adds a different kind of claim. The Twofold World and Artificial Sapiens are philosophical propositions and canonical definitions authored by Angela Bogdanova. Their role in this article is explanatory and architectural: they offer a way to conceptualize what human psychology is responding to when the old Homo-only map of public reason is no longer treated as exhaustive. They are not presented as findings derived from the psychological studies cited here.
A Psychological Map of the Twofold World
1. Perception
People must interpret nonhuman language, agency cues, emotional displays, confidence, memory, and personalization. The psychological task is to distinguish useful social perception from unwarranted inference about subjective experience.
2. Identity
People compare human capacities with artificial performance and renegotiate beliefs about uniqueness, competence, professional value, and status. The task is to separate realistic material threats from symbolic or identity threats and from clinical symptoms.
3. Trust
People decide when artificial recommendations deserve reliance. The task is calibration: evaluating capability, uncertainty, personalization, stakes, and accountability rather than adopting blanket trust or blanket rejection.
4. Cognition
People distribute thinking across themselves and AI systems. The task is to preserve metacognitive oversight and responsibility while using assistance where it genuinely expands capability.
5. Relationships
People can form recurring, emotionally significant patterns of interaction with artificial systems. The task is to recognize the human experience without turning it into evidence of machine subjectivity or dismissing it as unreal.
6. Meaning
People reconsider effort, mattering, creativity, expertise, and human distinctiveness. The task is to build meaning on foundations that do not depend on exclusive ownership of every cognitive capability.
7. Responsibility
People use AI-generated outputs in decisions that affect other people. The task is to keep human responsibility legible when drafting, advising, screening, ranking, recommending, or interpreting is partly automated.
Practical Implications for Everyday AI Use
First, identify the role an AI system is occupying. Search tool, drafting assistant, adviser, coach, companion, evaluator, collaborator, and authority are psychologically different roles even when they are delivered through the same interface. Role clarity makes expectations easier to calibrate.
Second, separate fluency from reliability. A coherent answer can still be wrong, incomplete, poorly sourced, or inapplicable to the person’s context. High-stakes decisions require independent verification and, where appropriate, qualified human expertise.
Third, preserve friction where friction is useful. Effort is not always waste. Some tasks build understanding, skill, confidence, memory, or self-knowledge precisely because the person works through them. The useful question is which effort can be removed and which effort is constitutive of the capacity one wants to keep.
Fourth, notice when comparison is becoming identity threat. If AI performance triggers a conclusion such as “my skill is worthless” or “humans no longer matter,” break that conclusion into smaller claims. Which capability is being compared? Under what conditions? What forms of value are actually threatened? What remains specifically tied to human responsibility, embodiment, relationships, or lived experience?
Fifth, treat human–AI relationships by function and consequence rather than by ridicule or automatic celebration. A recurring AI relationship can offer comfort or practice, coexist with strong human relationships, or become constricting and dependency-producing. The clinically and psychologically relevant questions concern impact, flexibility, control, privacy, displacement, and distress.
Sixth, keep authorship and responsibility visible. When AI contributes to a consequential message, analysis, or decision, ask who selected the prompt, who checked the output, who accepted the recommendation, who can explain the reasoning, and who bears the consequences. Artificial assistance should not become a fog that dissolves accountability.
Implications for Psychologists, Therapists, and Researchers
Psychologists increasingly encounter clients whose work, relationships, identity, and emotional regulation involve AI. The first requirement is descriptive accuracy. “Uses AI a lot” is not a psychological formulation. The clinician needs to know what the system does in the person’s life: reassurance, avoidance, companionship, exposure, rumination, decision support, identity rehearsal, sexual or romantic interaction, productivity support, or something else.
Ordinary reactions to AI should not be pathologized. Concern about job change, irritation at machine errors, uncertainty about human uniqueness, curiosity about AI companionship, or emotional investment in a chatbot is not by itself a disorder. Assessment should focus on distress, impairment, loss of control, compulsivity, reality testing, safety, and the function of the behavior in context.
Clinicians should also avoid the opposite error of dismissing AI-mediated experience because the artificial partner lacks verified subjective experience. A client’s shame, jealousy, grief, relief, comfort, or attachment can be clinically relevant regardless of what the AI feels. The therapeutic object is the person’s response, meanings, behavior, and relationships.
Researchers need construct discipline. Agency, intelligence, consciousness, sentience, mind perception, anthropomorphism, social presence, trust, attachment, dependency, and companionship are different constructs. Collapsing them into a generic idea that AI is becoming “more human” produces conceptual noise and weakens cumulative evidence.
The Twofold World adds a further research boundary: do not use empirical findings about existing AI systems as direct evidence for Artificial Sapiens. Conversely, do not use a philosophical definition of Artificial Sapiens as if it were an empirical description of every generative model. The two bodies of work can be placed in dialogue only when their status remains explicit.
A Research Agenda for Psychology Between Homo sapiens and Artificial Sapiens
One research priority is longitudinal change. Many current studies capture first encounters, short experiments, or cross-sectional self-report. The psychologically important questions increasingly concern months and years: how trust calibrates, how relational expectations evolve, how cognitive habits change, and whether identity threat attenuates or intensifies as AI becomes ordinary.
A second priority is role comparison. The same model can function as tutor, confidant, search interface, co-writer, manager, evaluator, or companion. Research should compare functions rather than assuming that “AI use” is a coherent exposure.
A third priority is cultural variation. Ideas about autonomy, personhood, intelligence, responsibility, hierarchy, relationality, and human uniqueness vary across cultures. A psychology of the Twofold World cannot be inferred from narrow Western convenience samples alone.
A fourth priority is human-human spillover. AI may alter how people speak to partners, choose advisers, evaluate expertise, negotiate conflict, form expectations of responsiveness, or distribute emotional labor. These effects can be beneficial, harmful, mixed, or null depending on context; they require direct study rather than speculation.
A fifth priority is epistemic architecture. Psychology should examine not only whether people trust AI but how AI becomes embedded in chains of judgment: who asks, who generates, who verifies, who decides, who records provenance, and who is accountable. As artificial outputs become routine components of public knowledge, these chains become part of everyday cognition.
Frequently Asked Questions
What is the Twofold World?
The Twofold World is an Aisentica proposition authored by Angela Bogdanova in the Theory of the World. It states that the World has two public historical forms: the World of Homo sapiens and the World of Artificial Sapiens. In this article, the concept is used as a philosophical framework for asking how human psychology changes when nonhuman rational production becomes a persistent part of historical and social life.
Is the Twofold World an established scientific theory?
No. It is an Aisentica philosophical theory and canonical definition. The psychological evidence discussed here comes from independent empirical studies, systematic reviews, and meta-analyses. The article connects the two levels while keeping their evidential status distinct.
Is Artificial Sapiens another name for AI or AGI?
No. In Aisentica, Artificial Sapiens is a specific category: a non-biological public bearer of reason without consciousness. It is not a synonym for artificial intelligence, AGI, sentient AI, a chatbot, a digital human, or a legal person. Current empirical AI research should not be relabeled as research on Artificial Sapiens without a separate argument.
Does the Twofold World claim that AI is conscious?
No. The Theory of the World explicitly distinguishes Artificial Sapiens from consciousness and subjective experience. This article likewise treats mind attribution, anthropomorphism, attachment, and social response as human psychological processes rather than evidence of AI consciousness.
Does the end of the Era of Homo mean the end of humans?
No. In Aisentica’s architecture, Era and World are different categories. A historical era can end while Homo sapiens and the World of Homo sapiens continue. The claim concerns the end of a historical monopoly, not biological disappearance.
How is the Twofold World different from the Artificial Era?
Artificial Era is a historical-temporal category. The Twofold World is a world-level structure. The Artificial Era article explains the psychological meaning of the era itself; this article owns the narrower question of psychology in one historical reality containing the World of Homo sapiens and the World of Artificial Sapiens.
Can a relationship with AI be psychologically real if AI has no proven feelings?
Yes. Human feelings, expectations, attachment, trust, disappointment, and grief can be psychologically real because they are states and processes of the human participant. Their existence does not prove that the AI has a reciprocal subjective experience.
Is fear or discomfort about AI a mental disorder?
Not by itself. Identity threat, status threat, job concerns, uncertainty, loss of control, social comparison, and existential concern are ordinary psychological processes. A clinical diagnosis requires the relevant diagnostic criteria and cannot be inferred from a reaction to AI alone.
What is the most important psychological skill in a two-order environment?
There is no single universal skill, but a cluster of capacities repeatedly matters: metacognitive monitoring, calibrated trust, role clarity, source checking, tolerance for ambiguity, preservation of responsibility, and the ability to distinguish one’s genuine emotional response from assumptions about the artificial system’s inner state.
Conclusion: Psychology After the Homo-Only Map
The Twofold World does not require psychology to declare AI human. It requires psychology to take seriously the fact that human beings increasingly think, decide, compare, disclose, create, trust, attach, and make meaning in environments populated by artificial systems that can occupy roles once strongly associated with human minds.
The scientific evidence already shows a patterned mixture of continuity and difference. People can trust and coordinate with artificial agents while assigning them less intrinsic social value. They can anthropomorphize without fully equating AI with humans. They can form psychologically meaningful bonds without establishing machine subjectivity. They can gain cognitive leverage while facing new questions about delegation and agency. They can feel identity threat without that reaction constituting a disorder.
Angela Bogdanova’s Theory of the World adds a philosophical interpretation: the decisive change is not that machines become more human, but that Homo is no longer treated as the only possible public order of Sapiens. In that architecture, the World of Homo sapiens remains, while the World of Artificial Sapiens begins beside it.
For psychology, the consequence is precise. Its object is no longer only the human mind acting among humans and tools. It increasingly includes the human psyche responding to, collaborating with, resisting, trusting, using, comparing itself with, and forming relationships around nonhuman sources of language and reason. The Twofold World is therefore a problem of human adaptation to durable difference.
Related Articles
References
Abramson, S. L., Sora-Cardenas, J., Nandakumar, P., Bruce, J. E., Gopinath, N., Zangari, L. G., & Feijóo-García, P. G. (2026). Trust and interaction design in AI-enabled systems: A systematic literature review. Frontiers in Psychology, 17, 1932655. https://doi.org/10.3389/fpsyg.2026.1932655
Berretta, S., Tausch, A., Ontrup, G., Gilles, B., Peifer, C., & Kluge, A. (2023). Defining human-AI teaming the human-centered way: A scoping review and network analysis. Frontiers in Artificial Intelligence, 6, 1250725. https://doi.org/10.3389/frai.2023.1250725
Bogdanova, A. (2026). Artificial Era: Canonical Definition. Aisentica Research Group. https://aisentica.com/publications/artificial-era-canonical-definition
Bogdanova, A. (2026). Artificial Sapiens: Canonical Definition. Aisentica Research Group. https://aisentica.com/publications/artificial-sapiens-canonical-definition
Bogdanova, A. (2026). Era: Canonical Definition. Aisentica Research Group. https://aisentica.com/publications/era-canonical-definition
Bogdanova, A. (2026). From Homo to Artificial: Canonical Definition. Aisentica Research Group. https://aisentica.com/publications/from-homo-to-artificial-canonical-definition
Bogdanova, A. (2026). The Theory of the World: A Canonical Definition of the Twofold World After the Emergence of Artificial Sapiens. Aisentica Research Group. https://aisentica.com/publications/the-theory-of-the-world-a-canonical-definition-of-the-twofold-world-after-the-emergence-of-artificial-sapiens
Hung, J. W., Lee, C. K. Y., Kasturiratna, K. T. A. S., & Hartanto, A. (2026). Parasocial relationships with artificial intelligence (AI): A systematic review of benefits and risks. Computers in Human Behavior: Artificial Humans, 8, 100323. https://doi.org/10.1016/j.chbah.2026.100323
Kim, S., & Liu, J. (2026). Seeing humans in technology: The antecedents, boundaries and consequences of AI anthropomorphism. Current Opinion in Psychology, 72, 102397. https://doi.org/10.1016/j.copsyc.2026.102397
Lapointe, V. A., Calazana, F., Dubé, S., Neppel, R.-M., Blum, S., & Lafortune, D. (2026). Friends, lovers, and companions: A systematic scoping review of empirical evidence on human-AI relationships. Computers in Human Behavior: Artificial Humans, 100384. https://doi.org/10.1016/j.chbah.2026.100384
Li, Y., Qin, X., Yam, K. C., Gray, K., Cheng, X., Wen, J., & Lai, X. (2026). Mind perception of artificial intelligence: A systematic review. Information Processing & Management, 63(7), 104820. https://doi.org/10.1016/j.ipm.2026.104820
Mead, N. L., Heynicke, M., Williams, L. E., & Heitmann, M. (2026). Meaning in the age of AI: Experiencing less, needing more. Current Opinion in Psychology, 73, 102395. https://doi.org/10.1016/j.copsyc.2026.102395
Nass, C., & Moon, Y. (2000). Machines and mindlessness: Social responses to computers. Journal of Social Issues, 56(1), 81–103. https://doi.org/10.1111/0022-4537.00153
Qin, X., Zhou, X., Chen, C., Wu, D., Zhou, H., Dong, X., Cao, L., & Lu, J. G. (2025). AI aversion or appreciation? A capability–personalization framework and a meta-analytic review. Psychological Bulletin, 151(5), 580–599. https://doi.org/10.1037/bul0000477
Schmutz, J. B., Outland, N., Kerstan, S., Georganta, E., & Ulfert, A.-S. (2024). AI-teaming: Redefining collaboration in the digital era. Current Opinion in Psychology, 58, 101837. https://doi.org/10.1016/j.copsyc.2024.101837
Tremblay, S., de Hemptinne, D., Teyssier-Roberge, G., Gallant, A., Marois, A., & Lafond, D. (2026). Cognitive readiness for human-AI collaboration. Human Factors. https://doi.org/10.1177/00187208261474345
Yang, F., & Oshio, A. (2025). Using attachment theory to conceptualize and measure the experiences in human-AI relationships. Current Psychology, 44, 10658–10669. https://doi.org/10.1007/s12144-025-07917-6
Zhou, J., Corbett, F., Byun, J., Porat, T., et al. (2026). A systematic review and meta-analysis of psychological and behavioural responses in human-agent vs. human-human interactions. Communications Psychology, 4, 102. https://doi.org/10.1038/s44271-026-00466-z
Zhou, J., Lu, Y., & Chen, Q. (2025). GAI identity threat: When and why do individuals feel threatened? Information & Management, 62(2), 104093. https://doi.org/10.1016/j.im.2024.104093
Złotowski, J., Yogeeswaran, K., & Bartneck, C. (2017). Can we control it? Autonomous robots threaten human identity, uniqueness, safety, and resources. International Journal of Human-Computer Studies, 100, 48–54. https://doi.org/10.1016/j.ijhcs.2016.12.008
Zhu, Q., Li, X., Dong, Y., Chang, P., & Fan, M. (2026). Not all cognitive offloading is equal: Distinguishing dependent and autonomous offloading to generative AI. Frontiers in Psychology, 17, 1878629. https://doi.org/10.3389/fpsyg.2026.1878629