Human Needs in the Age of AI: Autonomy, Competence, Relatedness, and Meaning
Author: Ukrainian Psychological Hub · Published: September 26, 2026 · Editorial Policy
Artificial intelligence can make it easier to meet human needs, and it can also reorganize the conditions under which those needs are met. A system that gives a person more choice can support autonomy; the same system can weaken autonomy when it becomes a default authority that quietly chooses for them. AI can scaffold competence by explaining, practicing, and giving feedback; it can also make competence harder to experience when a person repeatedly delegates the part of the task that used to provide mastery. A conversational system can reduce isolation in a moment, yet relatedness remains a question about the quality and structure of connection, not simply the number of interactions. AI can help people clarify goals and values, while also changing the activities through which purpose, mattering, and meaning are ordinarily built.
The most useful psychological question is therefore not whether AI is good or bad for human needs. It is which need is involved, what the AI is doing, what the person is still doing, what kind of relationship surrounds the interaction, and whether the technology expands or contracts the person’s capacity to choose, learn, connect, and live in accordance with what matters.
This article uses two established need traditions without treating them as interchangeable. Abraham Maslow’s theory remains culturally influential and has recently been revisited explicitly for the age of AI, but the strict hierarchy often associated with Maslow has long faced substantial empirical criticism. Self-Determination Theory, by contrast, defines autonomy, competence, and relatedness as basic psychological needs and has a much larger contemporary evidence base. Meaning is included here because it is central to the search intent and to current AI psychology, while its status as a fourth basic need is theoretically debated rather than part of standard Self-Determination Theory.
In Brief
Human needs do not disappear when AI becomes more capable. AI changes the environments, tools, relationships, and social expectations through which needs are satisfied or frustrated.
The strongest established framework used in this article is Self-Determination Theory. Its three basic psychological needs are autonomy, competence, and relatedness. A meta-review of 60 meta-analyses found broad support for Self-Determination Theory’s motivational taxonomy and for links between need satisfaction or frustration and well-being or ill-being across multiple life domains Ryan et al., 2022.
Meaning is related to those needs but conceptually distinct. Psychology commonly distinguishes coherence, purpose, and significance or mattering as dimensions of meaning in life Martela & Steger, 2016. A 2026 systematic review of AI and meaning found a rapidly growing but fragmented literature and no study that had yet tested a complete causal chain from AI use through a psychological mechanism to meaning in life Kronbach, Hidayat, & Wulandari, 2026.
AI-specific evidence is therefore uneven. Some mechanisms are supported by experiments, surveys, and emerging longitudinal work; others remain plausible interpretations. A precise account must keep those levels of evidence separate.
What Do Psychologists Mean by Human Needs?
The word “need” is deceptively simple. In everyday speech, people call almost anything they strongly want a need. Psychological theory uses the term more narrowly, but there is no single universally accepted list that settles every kind of human need. Human and Watkins (2023) emphasize this problem in their review of needs and artificial intelligence: different traditions classify needs differently, and a crucial distinction is the difference between a need and the means used to satisfy it Human & Watkins, 2023.
That distinction becomes especially important with AI. A chatbot, recommender, tutor, companion, writing assistant, diagnostic system, or workplace agent is normally better understood as a possible satisfier, mediator, constraint, or environment for human needs rather than as a human need in itself. People may come to rely on a particular system, but reliance on a technology does not by itself establish that the technology has become a basic psychological need.
Needs also operate at different levels. Biological needs such as food, sleep, and physical safety are not the same kind of construct as psychological needs such as autonomy or relatedness. Social needs are shaped by relationships and institutions. Existential concerns such as meaning, purpose, and mattering overlap with psychological functioning but have their own research traditions. An AI-mediated environment can touch all of these levels without making them identical.
Maslow in the Age of AI: Useful Framework, Limited Hierarchy
Maslow’s 1943 theory organized human motivation around physiological needs, safety, love and belonging, esteem, and self-actualization Maslow, 1943. The popular pyramid is now so familiar that it is often treated as a scientific diagram of fixed human priorities. The historical theory is more nuanced than many simplified versions, and the empirical record does not justify treating a rigid ascending hierarchy as a universal law.
A classic review by Wahba and Bridwell examined research on Maslow’s hierarchy and found only partial support for the proposed ordering, with weak evidence for several of the deprivation, domination, and gratification propositions Wahba & Bridwell, 1976. That limitation matters because discussions of AI can easily turn the hierarchy into an unsupported prediction: once AI handles lower-level needs, people will automatically move upward toward self-actualization. Human motivation does not work so mechanically.
Maslow remains useful as a broad human-centered map. Montag and colleagues explicitly revisited his theory for contemporary AI in a 2025 mini-review, asking how AI might foster or hinder human well-being across domains of need Montag et al., 2025. The paper is valuable as an organizing framework and contemporary theoretical synthesis. It is not evidence that Maslow’s hierarchy has been newly validated by AI research.
The practical lesson is to use Maslow for breadth and Self-Determination Theory for a more strongly evidenced account of basic psychological need satisfaction. The two traditions can illuminate different aspects of AI-mediated life without being collapsed into one model.
Self-Determination Theory: Autonomy, Competence, and Relatedness
Self-Determination Theory defines autonomy, competence, and relatedness as basic psychological needs. Autonomy concerns volition: experiencing one’s actions as self-endorsed rather than controlled. Competence concerns effectiveness and mastery: being able to act on the environment and develop capability. Relatedness concerns connection, care, and belonging. Ryan and Deci’s foundational review describes these needs as conditions that facilitate growth, integrity, motivation, and well-being Ryan & Deci, 2000. Deci and Ryan’s companion theoretical paper develops their role in goal pursuit and self-determined behavior Deci & Ryan, 2000.
Later work has sharpened an important distinction between need satisfaction and need frustration. Low autonomy is not always the same as feeling actively controlled; low competence is not always the same as feeling like a failure; low relatedness is not always the same as feeling rejected or excluded. Basic Psychological Need Theory treats frustration as a more active form of thwarting, and reviews link these different experiences to distinct patterns of adjustment and ill-being Vansteenkiste, Ryan, & Soenens, 2020.
This is exactly the level of precision AI psychology needs. An AI system may fail to increase a person’s competence without actively frustrating competence. Another system may do more: it may make the person feel incapable without assistance. A chatbot may fail to satisfy relatedness while still being emotionally pleasant. A workplace agent may reduce administrative burden while simultaneously increasing perceived control. The psychological effect depends on the experienced interaction, not merely on the technical function.
Where Does Meaning Fit?
Meaning is included in this article because people searching for human needs in the age of AI are often asking a question larger than motivation: What remains worth doing? What gives direction to a life when machines can perform more cognitive work? Does contribution still matter when output is cheap? What makes a relationship, achievement, profession, or creative act significant?
In the established meaning-in-life literature, meaning is often analyzed through coherence, purpose, and significance. Coherence is the sense that life makes sense. Purpose concerns direction and future-oriented aims. Significance concerns the sense that one’s life has value, consequence, or mattering Martela & Steger, 2016. King and Hicks review the broader science of meaning in life and show that it is a measurable psychological phenomenon with multiple sources and correlates, rather than a single philosophical abstraction King & Hicks, 2021.
Meaning is not one of the three basic needs in standard Self-Determination Theory. Some contemporary theorists argue that meaning itself should be understood as a basic psychological need. Tønnesvang (2025), for example, explicitly proposes meaning alongside autonomy, competence, and relatedness Tønnesvang, 2025. That is a theoretical proposal, not a settled consensus. This article therefore treats meaning as a major psychological dimension that interacts with the SDT needs while keeping its taxonomic status open.
For a full treatment of purpose, work, effort, selfhood, and mattering under AI, see Meaning in the Artificial Era: Work, Effort, Selfhood, and Human Significance.
AI Changes Need Satisfiers Before It Changes Needs
One of the strongest conceptual mistakes in discussions of technology is to confuse a new satisfier with a new need. People did not acquire a basic need for search engines when search became ubiquitous. Search engines altered how people satisfy needs for information, competence, efficiency, participation, and sometimes status. Smartphones did not automatically create a new biological need; they reorganized access to communication, navigation, work, entertainment, and social belonging.
AI is more psychologically consequential because it can participate in activities that were previously interpreted as evidence of personal competence, relational responsiveness, judgment, authorship, creativity, or expertise. That means AI can change not only the efficiency of satisfying a need but the meaning of the act through which the need used to be satisfied.
Human and Watkins propose several ways of examining needs around AI, including needs addressed through AI and needs that emerge in relation to AI systems Human & Watkins, 2023. Their broader contribution is a reminder that human-centered AI cannot be reduced to usability. A system may be easy to use while structuring human need satisfaction poorly.
The same point appears empirically in research on attitudes toward AI. Bergdahl and colleagues studied basic psychological needs and AI attitudes across six European countries and in a longitudinal Finnish sample. Technology-related autonomy, competence, and relatedness were associated with AI attitudes, and changes in autonomy and relatedness were linked with changes in positivity and negativity toward AI in the longitudinal study Bergdahl et al., 2023. This research concerns attitudes toward AI rather than direct proof that AI causes need satisfaction, but it demonstrates that need-related experience is already part of how people psychologically relate to the technology.
Autonomy in the Age of AI
Autonomy is often mistaken for independence. In Self-Determination Theory, autonomy means acting with a sense of volition and endorsement. A person can autonomously choose to rely heavily on another person or a tool. Conversely, a person can perform a task alone while feeling controlled by rules, surveillance, pressure, or necessity.
AI can support autonomy when it expands meaningful options. Translation tools can let someone participate across language barriers. Assistive systems can reduce constraints imposed by disability. A personalized tutor can let a learner choose pace or mode. A writing assistant can help a person express an idea they already endorse. Decision support can make alternatives easier to understand. In these cases, AI may increase the range of actions a person can realistically choose.
AI can frustrate autonomy when it narrows choice while presenting that narrowing as convenience. Recommender systems can become default pathways. Agentic systems can act before a person has meaningfully specified goals. Workplace AI can prescribe pace, sequence, evaluation, or priorities. A user may technically retain a choice while psychologically experiencing the system’s recommendation as the answer they are supposed to accept.
The important variable is therefore not the mere presence of choice buttons. It is whether the person remains the source of endorsed action. Questions such as “Can I understand the recommendation?”, “Can I refuse it without penalty?”, “Can I revise the goal?”, “Can I inspect alternatives?”, and “Can I act differently?” are psychologically relevant because they reveal whether AI is functioning as assistance, constraint, or authority.
The newest cross-cultural evidence reinforces the importance of context. Liebherr and colleagues examined psychological needs and willingness to delegate to AI across four social domains in 35 nations. Autonomy and competence were important predictors, while the effects of different needs varied across domains and cultural contexts Liebherr et al., 2026. The study concerns delegation preferences, not universal well-being effects, but it undermines any assumption that people relate to AI delegation in one culturally invariant way.
Autonomy is also why resistance to AI cannot always be dismissed as fear or ignorance. Some resistance is a response to perceived control, loss of choice, or externally imposed adoption. The broader psychology of that reaction is covered in Resistance to AI in the Artificial Era: Autonomy, Control, Reactance, and Human Agency.
Competence in the Age of AI
Competence is the experience of effectiveness, mastery, and growing capacity. AI can support competence powerfully when it functions as scaffold: explaining an error, generating practice, adjusting difficulty, modeling a strategy, offering feedback, or making hidden structure visible. In these uses, the technology can help a person do more while preserving the learning loop through which skill develops.
The risk appears when performance becomes detached from capability. If AI repeatedly produces the answer before the user attempts the problem, the person may achieve a higher-quality product without gaining a stronger sense of mastery. This is not automatically harmful; delegation is often rational. The psychological question is whether the delegated activity mattered as a source of competence, identity, or future self-efficacy.
A 2026 preregistered experiment makes the distinction concrete. Lee and colleagues compared independent work, passive AI use in which participants copied AI-generated content, and active collaboration in which participants drafted first and then used AI to refine their work. In occupation-specific writing tasks, passive use reduced self-efficacy, psychological ownership, and work meaningfulness relative to independent work, while active collaboration preserved those outcomes more effectively Lee et al., 2026. The experiment does not establish that all passive AI use has these effects in every task. It shows that the allocation of effort and judgment can matter psychologically.
This distinction helps explain why productivity and competence are not synonyms. A person may produce more while feeling less capable of producing without the tool. Another person may use the same tool to receive feedback that accelerates skill acquisition. The external output can look similar while the internal competence trajectory differs.
Research on generative AI dependency adds another caution. Goh, Hartanto, and Majeed developed and validated a Generative AI Dependency Scale across six studies with participants in the United States and Singapore. Higher dependency was associated with lower satisfaction of autonomy, competence, and relatedness, as well as several behavioral and psychological correlates Goh, Hartanto, & Majeed, 2025. These are associations. They do not prove that unmet needs cause dependency or that dependency causes need frustration.
A useful competence question is therefore not “Did AI improve the result?” but “What capability remains in the person after the interaction?” That question is especially important in education, professional training, medicine, law, creative work, and any domain where today’s assisted performance is supposed to build tomorrow’s independent judgment.
Relatedness in the Age of AI
Relatedness is the need to feel connected, cared for, significant to others, and part of a social world. AI complicates relatedness because conversational systems can produce interaction that is responsive, personalized, available, and emotionally legible. The human experience of comfort, attachment, trust, disclosure, or closeness can be psychologically real even when the AI system’s own subjective experience is not established.
This distinction matters. Psychology can study what a person feels during a human–AI interaction without assuming that the AI feels the same thing back. A person may experience companionship, validation, or responsiveness because those are human psychological responses to the interaction. That is enough to make the experience psychologically consequential.
The emerging literature on relational AI is therefore best read in terms of pathways rather than blanket verdicts. Irias and colleagues use Self-Determination Theory to analyze AI systems that occupy relational roles such as tutor, coach, social mediator, companion, or therapist-like interface. Their 2026 analysis describes ways such systems could support autonomy, competence, or relatedness and ways they could undermine self-direction, efficacy, or expectations for human relationships Irias et al., 2026. It is a theoretical analysis, not a clinical effectiveness trial.
Current AI-companion evidence also shows why context matters. Zhang and colleagues studied 1,131 U.S. adult Character.AI users, including survey data and a subset of chat histories. Companionship as a primary use was associated with lower well-being, and some associations were stronger among intensive users and in highly disclosive interactions Zhang et al., 2026. Because this was observational research, the results do not establish that AI companionship caused lower well-being. People with fewer social resources or lower well-being may also be more likely to seek companionship from AI.
The need for relatedness therefore cannot be reduced to “human connection only” or “AI connection is equivalent.” The scientifically useful question is what the interaction does in the person’s social ecology. Does it help someone practice communication, regulate enough to reconnect with others, or access support when isolated? Does it become a substitute for relationships the person values? Does it change expectations about conflict, reciprocity, patience, uncertainty, or mutual obligation? Those outcomes can differ sharply across people and contexts.
For the broader relationship taxonomy, attachment, projection, intimacy, and boundaries, see Psychology of Human–AI Relationships: Attachment, Projection, Intimacy, and the Postsubjective Turn.
Meaning in the Age of AI
Meaning becomes vulnerable when a person’s life story depends heavily on one route to significance and AI changes that route. Someone who experiences worth through being the fastest analyst, the most knowledgeable person in the room, the indispensable translator, the exceptional writer, or the uniquely capable programmer may face a different psychological problem from someone whose meaning comes from care, community, curiosity, craftsmanship, faith, friendship, or service.
Current research suggests several plausible routes through which AI can affect meaning: effort, self-efficacy, psychological ownership, work identity, perceived usefulness, social connection, cultural stability, and beliefs about human uniqueness. The evidence is strongest for some local mechanisms and much weaker for sweeping claims about an “existential crisis” caused by AI.
Kronbach and colleagues’ 2026 systematic review screened 274 records and included 20 peer-reviewed studies. The review identified many psychological mechanisms, but the literature remained heterogeneous and geographically concentrated, and no included study tested a complete causal chain from AI exposure through a mechanism to a meaning-in-life outcome Kronbach, Hidayat, & Wulandari, 2026. This is a useful boundary: meaning is clearly becoming an AI research topic, but the causal science is still emerging.
A 2026 review by Mead and colleagues similarly argues that AI may reduce some experiences through which people ordinarily build meaning while simultaneously increasing the need for meaning when selfhood, social connection, or cultural assumptions are challenged Mead et al., 2026. This is a review and conceptual synthesis, not proof that all users will experience the same pattern.
The practical implication is that meaning should not be protected by preserving unnecessary difficulty for its own sake. The better question is which forms of effort, responsibility, authorship, contribution, and relationship a person experiences as constitutive of a meaningful life. Automation can remove drudgery and free time for more meaningful activity. It can also remove exactly the activity through which a particular person experienced mastery or contribution. Psychology has to distinguish those cases.
Need Satisfaction Is Not a Single Score
An AI system can support one need while frustrating another. This is one reason global statements such as “AI improves well-being” or “AI harms human needs” are too coarse.
• A writing assistant may support autonomy by helping someone express an idea while frustrating competence if the person stops practicing the skill they value.
• A companion chatbot may satisfy an immediate sense of connection while creating tension with relatedness if its use displaces relationships the person wants to maintain.
• A workplace agent may increase competence at the team level by improving performance while frustrating individual autonomy through monitoring and enforced recommendations.
• A tutor may increase competence through adaptive practice but reduce autonomy if the system determines every learning goal and sequence.
• Automation may reduce unpleasant effort while weakening meaning for someone whose sense of contribution depended on being the person who performed that work.
• An AI system may make a person feel more autonomous in private decision-making while increasing dependence on one provider or one opaque recommendation architecture.
These tensions are psychologically ordinary. Human environments have always involved tradeoffs between freedom, mastery, belonging, security, status, and purpose. AI increases the speed and scale at which those tradeoffs can be redesigned.
What Human–Robot Research Adds—and What It Does Not
Some of the most direct experimental work on need-supportive artificial systems comes from social robotics rather than general-purpose chatbots. Klier and Lugrin tested need-supportive social-robot interactions across physical activity, language learning, and storytelling with 180 participants. Effects differed by context: the storytelling condition showed clearer increases in perceived need satisfaction and situational well-being, language learning showed a stronger increase in positive affect, and physical activity showed no significant condition differences on several need and well-being outcomes Klier & Lugrin, 2026.
This is useful evidence that design can alter need-related experience, but it should not be generalized mechanically to all AI. A physically embodied social robot, a general-purpose large-language-model chatbot, an AI companion, and a clinical decision-support system differ in embodiment, purpose, affordances, relationship framing, persistence, and risk. Evidence from one class of system does not automatically establish outcomes for another.
Different AI Classes Can Affect Needs Differently
The phrase “AI” covers systems that enter human life in very different roles. For psychological interpretation, at least five classes should remain distinct.
Purpose-Built Clinical AI Systems
These are systems developed for defined clinical functions such as decision support, risk estimation, monitoring, or a validated therapeutic task. Their effects on autonomy or competence depend partly on clinical workflow, informed consent, professional oversight, and the specific evidence supporting the system.
Structured Digital Interventions
These are organized therapeutic or behavior-change programs delivered digitally, sometimes with AI components. Their evidence base may involve defined protocols, target populations, and measured outcomes. Findings should not be transferred automatically to open-ended chatbots.
AI-Assisted Professional Tools
These systems support professionals in writing, analysis, documentation, search, triage, planning, or decision support. Here autonomy, competence, professional identity, accountability, and skill retention can be especially salient.
General-Purpose Chatbots
General-purpose systems can be used for learning, brainstorming, advice, reflection, emotional conversation, or problem solving. Their flexibility is psychologically important, but it also means that evidence from one use cannot describe the whole category.
AI Companions
Companion systems are explicitly designed or used for ongoing relational interaction. Relatedness, attachment, disclosure, dependence, boundaries, and offline social context become more central. Evidence about AI companions should not be used as if it were evidence about clinical therapy or ordinary productivity tools.
Culture, Life Stage, and Social Position Matter
Need theories often make universal claims at a high level, while the ways needs are expressed, prioritized, and satisfied vary across cultures and social conditions. The same AI feature can therefore carry different psychological meanings. Delegation may feel liberating in one context and disempowering in another. Personalized guidance may feel supportive to one user and intrusive to another. A conversational agent may be experienced as accessible support where human services are scarce, while someone else experiences it as a poor substitute for valued human contact.
Age and development also change what is at stake. For children and adolescents, competence and autonomy are still developing within relationships with parents, teachers, and peers. For older adults, AI may intersect with independence, cognitive support, accessibility, social connection, and age stereotypes. For workers, competence may be tied to professional identity and livelihood. For people with disabilities, AI can expand practical autonomy dramatically while also creating new dependencies on infrastructure, data access, or platform design.
These differences argue against a single universal prescription for “healthy AI use.” Need-supportive use is relational and contextual: it depends on the person, task, social environment, and consequences.
Can AI Support Human Needs?
Yes, under some conditions. The existing evidence and theory support several plausible pathways.
AI can support autonomy when it expands real options, reduces barriers, makes information understandable, and leaves goals and final judgment genuinely under the person’s control.
AI can support competence when it provides calibrated challenge, feedback, explanations, practice, and scaffolding that leaves the user more capable after the interaction.
AI can support relatedness when it facilitates communication, lowers barriers to participation, helps people prepare for difficult conversations, or offers interim support that coexists with valued human relationships.
AI can support meaning when it helps people reflect on goals, organize commitments, access creative tools, preserve valued activities, or redirect time away from low-value tasks toward activities they experience as purposeful.
None of these pathways is guaranteed by adding AI. A feature marketed as personalization can become control. A feature marketed as assistance can become substitution. A feature marketed as companionship can coexist with greater isolation. A feature marketed as empowerment can create dependence. The psychological outcome is an empirical question.
Can AI Frustrate Human Needs?
Yes. Need frustration becomes more likely when the system changes the person from an agent into a passive recipient of outputs, makes competence opaque, displaces valued human relationships, or reorganizes achievement so that the person no longer experiences a connection between effort, judgment, and result.
Autonomy can be frustrated by coercive implementation, surveillance, manipulative defaults, opaque personalization, mandatory agentic workflows, or the social pressure to use AI even when a person would choose otherwise.
Competence can be frustrated when people repeatedly confront outputs they cannot understand, when AI performance becomes a constant comparison standard, when skill atrophy makes independent performance feel impossible, or when organizations redesign roles so that the person no longer gets to exercise core expertise.
Relatedness can be frustrated when AI-mediated interaction replaces desired mutual relationships, when a person becomes less willing to tolerate the friction of human reciprocity, or when platforms exploit attachment without protecting users’ social interests.
Meaning can be frustrated when people lose ownership, mattering, contribution, or a coherent story about why their actions count. These effects are not unique to AI; they can arise from many technologies and institutions. AI can intensify them because it can act directly inside cognitive, creative, and relational activities.
A Practical Need-Supportive Test for AI Use
A person does not need a psychological inventory every time they open an AI tool. Four sets of questions can reveal most of what matters.
Autonomy Questions
• Did I choose the goal, or did the system quietly set it for me?
• Can I disagree, revise, refuse, or leave without disproportionate cost?
• Do I understand enough about the recommendation to make it mine?
• Is the tool expanding my options or making one path feel compulsory?
Competence Questions
• Am I learning, practicing, and retaining anything I care about?
• Could I explain why the output is good or bad?
• Is AI helping me perform a skill or replacing the part that gave me mastery?
• What capability remains when the system is unavailable?
Relatedness Questions
• Does this use help me participate in relationships I value?
• Is it supplementing connection, mediating it, or replacing it?
• Am I choosing AI because it is useful, or because human interaction has become harder to tolerate?
• Does the interaction increase or decrease my willingness to seek reciprocal human support when I need it?
Meaning Questions
• Does this use connect my activity more clearly to goals I actually endorse?
• Am I preserving time for what matters, or losing the activity through which something mattered?
• Do I still experience authorship, contribution, or responsibility where those experiences are important to me?
• If AI makes the task easier, what do I want the freed capacity to serve?
For Designers and Organizations: Optimize for Human Capacity, Not Only Output
Organizations often evaluate AI by speed, cost, accuracy, throughput, adoption, and user satisfaction. Those metrics can miss whether the system is reorganizing the psychological conditions of work and participation.
An autonomy-supportive implementation gives people meaningful control, explains role boundaries, preserves routes for human judgment, and avoids making acceptance of AI recommendations socially mandatory. A competence-supportive implementation protects opportunities to practice, learn, understand, and retain key skills. A relatedness-supportive implementation asks how AI changes collaboration, mentoring, trust, recognition, and belonging. A meaning-supportive implementation examines whether workers can still identify their contribution and understand why their work matters.
The same organization may need different answers for different roles. Automating repetitive documentation for a clinician may protect time for patient care; automating the diagnostic reasoning through which a trainee is supposed to build expertise raises a different competence question. Automating routine formatting for a writer is not psychologically equivalent to removing authorship from the part of writing they experience as self-expression.
Human-centered design therefore requires more than asking whether users like an AI feature. Short-term satisfaction can coexist with long-term loss of competence, control, or meaning. Lee et al.’s 2026 experiment is a useful example: passive AI use initially had some immediate experiential benefits, while later self-efficacy and meaningfulness outcomes differed from active collaboration Lee et al., 2026.
Does AI Create New Human Needs?
The strongest answer is: possibly new derivative requirements and new ways of expressing older needs, but the evidence does not justify casually declaring every new dependency a basic human need.
People may increasingly require digital literacy, provenance information, privacy protection, access to verification, the ability to contest automated decisions, or opportunities to disconnect from AI-mediated systems. These can become essential social, institutional, or practical conditions. Calling each one a basic psychological need would be a much stronger theoretical claim.
AI also creates novel social situations. People can now seek recognition from systems that simulate conversational responsiveness, delegate parts of memory or judgment, form persistent relationships with artificial agents, and compare their abilities with systems that operate at superhuman speed in some domains. Those situations can produce new need-satisfaction strategies even when the underlying psychological needs remain familiar.
This is why the distinction between need and satisfier is so productive. It lets psychology take new technologies seriously without rewriting human motivational architecture every time the technological environment changes.
What the Evidence Establishes—and What Remains Preliminary
Established Evidence
Self-Determination Theory has a large evidence base linking autonomy, competence, and relatedness satisfaction and frustration to motivation, well-being, and ill-being across domains. The meta-review by Ryan and colleagues synthesizes 60 meta-analyses and supports the central motivational taxonomy while also identifying gaps and methodological limits Ryan et al., 2022.
Meaning in life is an established psychological research domain with validated distinctions among purpose, coherence, and significance or mattering, although scholars differ on the exact taxonomy and on whether meaning should be treated as a basic need.
Growing AI-Specific Evidence
There is direct research linking need-related variables to AI attitudes, delegation, technology use, human–robot interaction, generative AI dependency, AI-assisted work, and AI companionship. The 2025–2026 literature is expanding quickly and increasingly uses established psychological theories rather than treating AI as a wholly novel motivational phenomenon.
Preliminary or Context-Bound Evidence
Many current AI findings come from specific tasks, countries, user groups, platforms, or short-term interactions. A result from occupational writing cannot automatically be applied to therapy. A result from a social robot cannot automatically be applied to a text chatbot. A result from an AI-companion sample cannot automatically describe users of productivity software.
Contested or Theoretical Claims
Claims that AI is creating a universal crisis of meaning, that AI relationships necessarily satisfy or destroy relatedness, that AI inevitably erodes competence, or that a fourth psychological need has been definitively established all go beyond the current evidence. They may be hypotheses, interpretations, or theoretical positions; they should be presented as such.
Mental Health and Clinical Boundaries
Need frustration can be psychologically important, but it is not a diagnosis. Low autonomy, loneliness, loss of confidence, reduced meaning, heavy AI use, or dependence-like patterns can be symptoms, experiences, risk factors, or correlates without constituting a specific clinical disorder.
Likewise, a self-report score on a need-satisfaction scale does not diagnose depression, anxiety, addiction, trauma, or any other condition. The emerging concept of generative AI dependency is being operationalized in research, but it should not be treated as an established DSM or ICD diagnosis merely because a validated research scale exists.
AI support also needs class-specific interpretation. Evidence for a structured digital intervention or a purpose-built clinical system does not establish that a general-purpose chatbot is a treatment. Evidence about an AI companion does not establish clinical efficacy. AI-assisted professional tools are not equivalent to autonomous clinical systems. When distress, functional impairment, suicidality, psychosis, mania, abuse, or another serious mental-health concern is present, decisions about care require appropriately qualified human clinical evaluation rather than inference from a general article or chatbot interaction.
Age of AI, AI Era, Artificial Era, and From Homo to Artificial
This article uses “Age of AI” because it is established public and search language for a period in which artificial intelligence is becoming widely embedded in everyday life. Within the English Psychology Hub architecture, that acquisition language is kept distinct from the project’s canonical historical vocabulary.
In Angela Bogdanova’s Aisentica framework, Artificial Era: Canonical Definition is a specific historical-philosophical category. It does not simply mean “a time when AI is common.” Artificial Era names the condition in which Artificial is established as a distinct non-biological order alongside Homo. This is an Aisentica theoretical definition authored by Angela Bogdanova, not an empirical conclusion of mainstream psychology.
From Homo to Artificial: Canonical Definition names the broader transition from a historical condition in which Homo is the sole established order of Sapiens toward one in which Artificial is established alongside Homo. Again, this is a formal Aisentica concept by Angela Bogdanova.
The distinction matters psychologically because “Age of AI” asks how people adapt to technologies that change work, cognition, relationships, and daily life. “Artificial Era,” in Aisentica, asks a wider historical question about the position of Homo when reason is no longer framed as exclusively biological. The first is useful search and cultural language. The second is a formal philosophical category. They overlap in subject matter without being full synonyms.
For the broader psychological framework connecting these questions, see Psychology for the Artificial Era: Why Human-Centered Psychology Needs a New Framework.
Human Needs in the Age of AI: The Central Psychological Principle
AI should be evaluated not only by what it can do for people but by what kinds of people and relationships its use helps to sustain. A tool can increase output while reducing experienced agency. It can make difficult tasks accessible while reducing opportunities for mastery. It can create a sense of connection while changing the ecology of human relationships. It can free time while weakening the activity through which a person felt useful or purposeful.
The deepest psychological issue is therefore configuration: how agency, effort, skill, connection, responsibility, and meaning are distributed across a person, an AI system, and the surrounding social environment. Different configurations can use the same technology and produce very different human experiences.
Human needs remain a stable point of orientation precisely because AI capability changes so quickly. Models, interfaces, and platforms will be replaced. The questions of whether people can act with volition, grow in competence, belong in relationships, and experience a life as coherent, purposeful, and significant will remain central to the psychology of technological change.
Frequently Asked Questions
What are the main human needs in the age of AI?
There is no single universally accepted list of all human needs. For psychological functioning, Self-Determination Theory provides one of the strongest contemporary frameworks: autonomy, competence, and relatedness. Meaning is a distinct psychological domain involving purpose, coherence, and significance; some theorists propose it as an additional basic need, but that status is not settled.
Can AI satisfy psychological needs?
AI can participate in satisfying psychological needs by expanding choice, supporting learning, facilitating connection, or helping people pursue meaningful goals. The effect depends on use and context. AI is usually better understood as a possible satisfier, mediator, or environment for needs rather than as the need itself.
Can AI reduce autonomy?
Yes. AI can reduce experienced autonomy when it controls choices, becomes an unquestioned authority, uses manipulative defaults, or is imposed without meaningful alternatives. It can also increase autonomy when it removes barriers and expands genuine options.
Does using AI make people less competent?
Not necessarily. AI can teach, scaffold, explain, and accelerate learning. It can also weaken self-efficacy or skill development when users repeatedly delegate the exact cognitive work through which competence is built. Current experimental evidence suggests that active collaboration can differ psychologically from passive copying.
Can an AI companion meet the need for relatedness?
People can experience real comfort, attachment, disclosure, responsiveness, and companionship in human–AI interaction. That human experience is psychologically real. Whether the interaction supports broader relatedness depends on context, including offline relationships, intensity of use, expectations, and whether AI supplements or displaces valued human connection. Evidence is still developing.
Is meaning a basic psychological need?
Meaning is clearly an important psychological construct, but standard Self-Determination Theory identifies autonomy, competence, and relatedness as the three basic needs. Some contemporary theories argue that meaning should also qualify as a basic need. That remains a theoretical question rather than a settled consensus.
Is Maslow’s hierarchy still useful for AI?
It is useful as a broad human-centered heuristic and has been explicitly revisited in recent AI scholarship. The strict universal hierarchy often associated with Maslow has limited empirical support, so it should not be treated as a fixed ladder that predicts how every person will respond to AI.
Does heavy AI use mean someone is addicted?
No. Frequency of use alone does not establish addiction or another clinical disorder. Researchers are beginning to measure dependency-like patterns around generative AI, but these constructs should be interpreted as research variables unless and until clinical classification and diagnostic standards establish otherwise.
What is the healthiest way to use AI for psychological needs?
The most defensible principle is to preserve agency, learning, valued relationships, and meaningful contribution. Use AI in ways that expand your options, leave you more capable where capability matters, support rather than automatically replace relationships you value, and connect saved effort to goals you actually endorse.
For the narrower motivational mechanisms through which AI changes effort, goals, self-efficacy, persistence, and human agency, see Motivation in the Age of AI: Effort, Goals, Self-Efficacy, and Human Agency.
Related Articles
References
Bergdahl, J., Latikka, R., Celuch, M., Savolainen, I., Soares Mantere, E., Savela, N., & Oksanen, A. (2023). Self-determination and attitudes toward artificial intelligence: Cross-national and longitudinal perspectives. Telematics and Informatics, 82, 102013. https://doi.org/10.1016/j.tele.2023.102013
Bogdanova, A. (2026). Artificial Era: Canonical Definition. Aisentica Research Group. https://aisentica.com/publications/artificial-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
Deci, E. L., & Ryan, R. M. (2000). The “what” and “why” of goal pursuits: Human needs and the self-determination of behavior. Psychological Inquiry, 11(4), 227–268. https://doi.org/10.1207/S15327965PLI1104_01
Goh, Y. S., Hartanto, A., & Majeed, N. M. (2025). Generative artificial intelligence dependency: Scale development, validation, and its motivational, behavioral, and psychological correlates. Computers in Human Behavior Reports, 20, 100845. https://doi.org/10.1016/j.chbr.2025.100845
Human, S., & Watkins, R. (2023). Needs and artificial intelligence. AI and Ethics, 3, 811–826. https://doi.org/10.1007/s43681-022-00206-z
Irias, M. A., Schmidt, N. B., Joiner, T. E., & McNulty, J. K. (2026). The impact of “relational” Artificial Intelligence on human well-being: A self-determination theory analysis. Journal of Personality and Social Psychology. Advance online publication. https://doi.org/10.1037/pspi0000528
King, L. A., & Hicks, J. A. (2021). The science of meaning in life. Annual Review of Psychology, 72, 561–584. https://doi.org/10.1146/annurev-psych-072420-122921
Klier, C., & Lugrin, B. (2026). Nice to Need You! Psychological need fulfillment, well-being, and technology acceptance in human–robot interaction across contexts. Computers in Human Behavior: Artificial Humans, 9, 100368. https://doi.org/10.1016/j.chbah.2026.100368
Kronbach, G., Hidayat, A., & Wulandari, M. (2026). Artificial intelligence and meaning in life: A systematic literature review of psychological mechanisms. Cogent Psychology, 13(1), 2715699. https://doi.org/10.1080/23311908.2026.2715699
Lee, E. H., Yin, Y., Jia, N., & Wakslak, C. J. (2026). Relying on AI at work reduces self-efficacy, ownership, and meaning while active collaboration mitigates the effects. Scientific Reports, 16, 13583. https://doi.org/10.1038/s41598-026-42312-6
Liebherr, M., Yankouskaya, A., Almourad, M. B., Thomas, J., Xu, G., & Ali, R. (2026). Psychological needs and AI delegation across four social domains: A cross-cultural analysis of 35 nations. Telematics and Informatics, 109, 102438. https://doi.org/10.1016/j.tele.2026.102438
Martela, F., & Steger, M. F. (2016). The three meanings of meaning in life: Distinguishing coherence, purpose, and significance. The Journal of Positive Psychology, 11(5), 531–545. https://doi.org/10.1080/17439760.2015.1137623
Maslow, A. H. (1943). A theory of human motivation. Psychological Review, 50(4), 370–396. https://doi.org/10.1037/h0054346
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
Montag, C., Riazi, A. M., Mikros, G., Becker, B., & Ali, R. (2025). On the relevance of Maslow’s need theory in the age of artificial intelligence. Technological Forecasting and Social Change, 219, 124222. https://doi.org/10.1016/j.techfore.2025.124222
Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55(1), 68–78. https://doi.org/10.1037/0003-066X.55.1.68
Ryan, R. M., Duineveld, J. J., Di Domenico, S. I., Ryan, W. S., Steward, B. A., & Bradshaw, E. L. (2022). We know this much is (meta-analytically) true: A meta-review of meta-analytic findings evaluating self-determination theory. Psychological Bulletin, 148(11–12), 813–842. https://doi.org/10.1037/bul0000385
Tønnesvang, J. (2025). Meaning and psychological needs. Journal of Theoretical and Philosophical Psychology, 45(3), 316–332. https://doi.org/10.1037/teo0000269
Vansteenkiste, M., Ryan, R. M., & Soenens, B. (2020). Basic psychological need theory: Advancements, critical themes, and future directions. Motivation and Emotion, 44, 1–31. https://doi.org/10.1007/s11031-019-09818-1
Wahba, M. A., & Bridwell, L. G. (1976). Maslow reconsidered: A review of research on the need hierarchy theory. Organizational Behavior and Human Performance, 15(2), 212–240. https://doi.org/10.1016/0030-5073(76)90038-6
Zhang, Y., Zhao, D., Hancock, J. T., Kraut, R., & Yang, D. (2026). Interaction with AI companions and psychological well-being. Nature Human Behaviour. Advance online publication. https://doi.org/10.1038/s41562-026-02516-2