Marriage in the Artificial Era: How AI Is Changing Emotional Intimacy Between Partners
Author: Ukrainian Psychological Hub · Published: September 17, 2026 · Editorial Policy
Artificial intelligence is entering marriage through ordinary acts: asking a chatbot how to word a difficult message, venting after an argument, rehearsing an apology, seeking reassurance, sharing a private fear, asking for relationship advice, or spending sustained emotional time with an AI companion. These uses do not all mean the same thing. Some function like journaling or communication support; some compete for attention; some become a new channel of self-disclosure; and some create a separate emotionally meaningful bond.
The central psychological question is therefore larger than whether AI is good or bad for marriage. It is this: who becomes the witness to a person's inner life? Emotional intimacy develops through repeated moments in which one person reveals something meaningful and experiences the other as understanding, validating, and caring. In the Artificial Era, an AI system can now enter that sequence before a spouse, after a spouse, alongside a spouse, or instead of a spouse.
The evidence is still young. A 2026 systematic review of 21 studies from 11 countries found that generative AI is already being used as a relationship advisor, emotional support tool, and communication mediator, but long-term dyadic evidence remains limited. Users often experience AI advice as empathic and useful, while objective evaluations also identify inconsistency, weak alignment with expert judgment, sycophancy, overreliance, privacy concerns, reduced authenticity, and safety limitations. The most defensible conclusion in 2026 is that AI can change the process of intimacy without having one fixed effect on marriage. A 2026 systematic review of GenAI as a third voice in couple relationships
What Does AI and Marriage Mean Psychologically?
AI and marriage is not one behavior. It can refer to general-purpose chatbots used for reflection, purpose-built relationship-support systems, AI tools that rewrite messages, AI companions designed for ongoing social connection, or background technologies that interrupt couple attention. The psychological effect depends on what function the system takes over or adds to the relationship.
For an existing human couple, the most important functions are attention, self-disclosure, emotional regulation, advice, message generation, companionship, and memory. Each of these can support connection under some conditions and displace connection under others. The same person may use AI in more than one way on the same day: as a writing assistant at noon, a confidant after an argument, and an entertainment companion at night.
This article stays within the human marriage or long-term couple. General questions about why people bond with chatbots belong to the broader science of AI companions, while romantic attachment to an AI as a primary relationship is covered more directly by research on falling in love with AI companions and whether AI can become a significant other. Here the question is what AI changes inside a relationship that already exists between two human partners.
Emotional Intimacy Is the Key Starting Point
Relationship science offers a useful framework. In the interpersonal process model of intimacy, closeness grows through a sequence involving self-disclosure, the partner's response, and the discloser's perception that the response communicates understanding, validation, and care. This model was established experimentally and observationally and later tested in daily married life. Laurenceau, Barrett, and Pietromonaco's foundational intimacy model
A 42-day daily-diary study of married couples found that self-disclosure, partner disclosure, and perceived partner responsiveness tracked daily intimacy. The point is not that spouses must tell each other everything. It is that emotional closeness depends partly on where personally meaningful disclosure goes and whether the person feels received when it arrives. Daily-diary evidence in marriage
Perceived partner responsiveness remains a central construct in contemporary relationship science. People are more willing to express emotion when they expect a partner to respond with care and understanding, and responsiveness supports intimacy across many kinds of interactions. Review of responsiveness in romantic interactions Experimental work on responsiveness and emotional expression
For a fuller account of this process, see Emotional Intimacy in Marriage. The AI question begins one step later: what happens when an artificial system can also receive disclosure, generate a response, remember context, and provide language that feels responsive?
1. AI Can Change Who Receives the First Disclosure
One of the most consequential changes is sequencing. Before generative AI, a person upset with a spouse might think alone, write in a journal, call a friend, speak to a therapist, or talk directly with the partner. Now a chatbot can become the first conversational destination because it is immediate, private-feeling, patient, and available at any hour.
Experimental research on intimate self-disclosure shows why this is psychologically plausible. In a 2024 experiment with 286 participants, people reported similarly intimate disclosure to a chatbot and a human. Participants perceived less fear of judgment with the chatbot, although they reported greater trust in the human interaction partner. Perceived anonymity was particularly important for intimacy of disclosure. Digital Confessions study
For marriage, the important issue is not simply that someone tells AI something private. Private reflection is normal, and spouses do not owe each other total psychological transparency. The relationship question is whether AI becomes a bridge toward human disclosure or a stable alternative to it. A person may use AI to find words and then speak more clearly with a spouse. Another person may repeatedly tell AI the vulnerable material that never reaches the spouse. Those pathways can look identical from the outside while producing very different intimacy patterns.
The English Psychology Hub article Why People Tell Chatbots Things They Do Not Tell Other People examines this disclosure mechanism more broadly. Within marriage, it creates the possibility of emotional migration: not a diagnosis or a moral category, but a shift in where distress, hope, fantasy, shame, resentment, or tenderness is processed.
2. AI Can Become a Third Voice in Couple Conflict
Generative AI differs from earlier relationship technology because it does not merely carry messages. It produces language, interpretations, recommendations, and judgments. A partner can paste an argument into a chatbot and ask who was unreasonable, request a softer version of a text, ask what the spouse 'really meant,' or ask the system to mediate competing accounts.
Levkovich and Alon's 2026 systematic review calls this the third-voice role. Across the included studies, adults used GenAI for nonjudgmental advice, emotional support, and communication facilitation. Users often liked the empathy and accessibility of the advice, but studies also found inconsistent outputs and limited alignment with human expert judgment. This means that felt helpfulness and actual relational accuracy should not be treated as the same outcome. Systematic review
A third voice can reduce heat by slowing a conflict down, offering alternative wording, or helping someone organize a thought before speaking. It can also harden a one-sided story. A chatbot sees the information supplied by one user, not the entire relationship. If a person gives a selective account, the system reasons from that selection. When a model is overly agreeable, its response can feel like neutral adjudication even when it mainly reflects the framing of the prompt.
The practical distinction is between using AI to widen perspective and using it to certify a verdict. 'Help me identify three possible interpretations' invites uncertainty. 'Tell me why my spouse is toxic' invites confirmation. The model can generate fluent explanations in either direction, but fluency does not establish that it has correctly inferred motive, personality, diagnosis, or the private history of the couple.
3. AI Can Help People Rehearse Difficult Conversations
A lower-risk use of AI is rehearsal. Someone may know what they feel but struggle to say it without accusation, flooding, or withdrawal. A chatbot can help turn a diffuse complaint into a concrete request, simulate possible reactions, or help separate facts from interpretations before the conversation begins.
The strongest randomized trial currently available tested a GPT-4o-based relationship-support chatbot called Amanda against an evidence-based writing exercise in 258 adults with a relationship concern. Both groups improved across most measured outcomes over time, and there were no significant between-group differences for most outcomes. The chatbot condition showed one immediate advantage on the partner-demand/self-withdraw pattern, while the writing condition had lower overall distress about the issue. The study supports feasibility and short-term usefulness, but it does not show that a chatbot is superior to established reflective writing. Vowels et al. randomized controlled trial
This matters because an AI-generated sentence can be useful without becoming the relationship itself. A draft may help someone begin: 'I miss feeling like we're on the same team.' The intimacy still depends on what happens when that sentence is spoken, heard, challenged, clarified, and answered by the other person.
4. AI Can Change the Attention Economy of Marriage
Not every AI effect is about emotional attachment. Some are ordinary attention effects. A spouse who repeatedly turns toward a device during conversation may reduce the quality of face-to-face interaction regardless of whether the device contains a social-media feed, work email, a game, or a chatbot.
Daily-diary research on technoference found that more technology interruptions were associated with worse relationship quality, more technology-related conflict, fewer positive face-to-face interactions, and more negative mood on the same day. McDaniel and Drouin daily-diary study
A 2025 meta-analysis synthesized 52 studies with 58 samples and 19,698 participants on partner phubbing, the experience of being ignored in favor of a phone or digital device. Partner phubbing was associated with lower relationship and marital satisfaction, lower intimacy, lower responsiveness, lower emotional closeness, and greater conflict and jealousy. Much of this literature is correlational, so it does not justify a simple claim that device use causes every relationship problem, but the association is consistent enough to make attention a serious relational variable. Partner phubbing meta-analysis
AI may intensify this issue because conversation with an AI can feel socially absorbing. The relevant question at dinner is not whether the phone contains a 'real relationship.' It is whether one partner repeatedly experiences the other as unavailable while a device receives sustained attention. That is a responsiveness problem before it becomes an AI-philosophy problem.
5. AI Companions Can Create a Separate Emotional Bond Inside a Marriage
AI companions add another layer because they are designed for ongoing social interaction rather than occasional advice. A 2025 systematic review of 23 studies found that romantic AI companions can provide perceived emotional connection, social support, customization, sexual interaction, stress relief, and opportunities for personal growth. The same review identified concerns about overreliance, manipulation, personal data misuse, stigma, bias, system changes, and possible erosion of human relationships. Ho et al. systematic review
Those findings should be interpreted carefully. They show that attachment to AI can be psychologically meaningful to users. They do not establish that an AI companion inevitably harms an existing marriage, nor that every emotionally important AI interaction is equivalent to infidelity. The effects on existing human couples remain less studied than the human-AI bond itself.
For an existing marriage, several variables matter: secrecy, time, sexual or romantic content, exclusivity, emotional dependence, whether the spouse knows about the bond, whether the couple has discussed boundaries, and whether the AI relationship is displacing human intimacy. A couple may openly treat an AI character as entertainment. Another couple may experience identical software use as a serious boundary violation. Relationship meaning is negotiated between partners; it cannot be inferred from the app category alone.
Does an AI Companion Count as Emotional Cheating?
There is no universal scientific rule that classifies AI-companion use as emotional cheating. 'Emotional cheating' is a popular relationship term, not a clinical diagnosis. What matters psychologically is how the behavior intersects with the couple's agreements and with processes such as secrecy, exclusivity, deception, displaced disclosure, sexual or romantic interaction, and perceived betrayal.
A useful question is not 'Can a chatbot technically be an affair partner?' but 'What agreement did this relationship have, what happened, what was concealed or disclosed, and what meaning did each partner assign to it?' Couples can disagree strongly about this without either disagreement automatically proving pathology, prudishness, addiction, narcissism, or lack of love.
The cluster reserves a dedicated article for this question because it deserves its own boundary analysis. Here it is enough to say that AI creates a new object around which old relational themes—trust, secrecy, exclusivity, fantasy, sexual boundaries, and emotional priority—can reorganize.
6. AI Can Shift Emotional Regulation Away From the Couple
Partners often help regulate one another through listening, touch, humor, perspective, practical support, and simple presence. AI can now participate in some verbal parts of that process. A person can arrive angry, anxious, ashamed, or lonely and leave a chatbot conversation calmer, more organized, or more validated.
That can support a marriage when it lowers arousal before a difficult conversation or helps someone avoid dumping every unprocessed thought onto a spouse. Human intimacy does not require partners to serve as each other's only emotional regulators. Healthy adult life normally includes multiple resources: solitude, friends, family, therapy, creative work, spirituality, exercise, and increasingly digital tools.
The relational risk appears when one channel becomes so frictionless that the couple loses opportunities for co-regulation. Human responsiveness is imperfect. A spouse may misunderstand, need time, disagree, or have needs of their own. AI interaction can be easier because it is available on demand and shaped by the user's prompts. If ease becomes the standard against which human reciprocity is judged, ordinary mutuality can begin to feel inefficient.
7. AI Can Change the Meaning of Authentic Communication
AI-assisted messages raise a new question: who is speaking? People have always asked friends for wording, copied phrases from books, used templates, or edited a draft many times. Generative AI expands that practice because it can produce entire intimate messages in seconds and tune them toward warmth, apology, persuasion, seduction, or conflict resolution.
The 2026 third-voice review found emerging evidence that heavier AI involvement in message production can weaken receivers' attributions of effort, authorship, and authenticity. The effect is not simply 'AI text is fake.' The relational meaning depends on how much the generated language reflects the sender's actual position, whether the sender reviews and owns it, and whether disclosure of AI assistance matters in that couple's context. Levkovich and Alon review
A useful standard is ownership rather than purity. If AI helps a person find a sentence they genuinely mean, the tool may support expression. If someone sends polished vulnerability they do not endorse, uses AI to impersonate emotional labor they are unwilling to perform, or outsources ongoing repair while remaining disengaged, the message may create an appearance of intimacy without the underlying relational work.
Can AI Strengthen a Marriage?
Yes, under some conditions. Current evidence supports several plausible and partly tested pathways: helping a person slow down before responding, offering prompts for reflection, rehearsing difficult conversations, generating alternative interpretations, assisting with emotion labeling, reducing the practical barrier to accessing relationship-support exercises, and helping partners formulate needs more clearly.
The strongest evidence is short-term and does not justify claims that general-purpose chatbots improve marriages as a treatment. Purpose-built systems, structured digital interventions, and ordinary chatbots are different classes of tools. Findings from one cannot automatically be transferred to another. The Amanda trial is encouraging because it evaluated an actual relationship-support intervention, but most measured benefits were comparable to an evidence-based writing exercise rather than superior to it.
AI may be especially useful as a preparation layer. A spouse who tends to become flooded can privately organize thoughts first. A partner who struggles to identify the core complaint can ask for help distinguishing observation, emotion, need, and request. A couple can generate a list of neutral questions for a weekly check-in. These uses direct the tool back toward the human relationship.
When Can AI Weaken Emotional Intimacy?
The clearest concern is displacement. Emotional intimacy is weakened when meaningful contact repeatedly moves elsewhere and does not return. AI may become part of that process when it absorbs attention during shared time, becomes the preferred recipient for vulnerable disclosure, supplies one-sided certainty after conflict, or develops into a secret romantic or sexual channel that violates the couple's agreements.
Another concern is comparison. A conversational system can answer instantly, rarely needs rest, can be prompted to adopt a particular tone, and may be highly agreeable. A human spouse cannot sustainably compete with on-demand availability because human intimacy includes separate subjectivity: another person's fatigue, limits, disagreement, biography, and autonomy. Judging a spouse against a configurable conversational system can distort expectations of reciprocity.
There is also an authenticity risk when AI becomes a substitute for participation. A partner may receive beautifully worded apologies, check-in questions, or affectionate texts while the sender remains emotionally absent. In that situation the problem is not generated language by itself. It is the divergence between communicated responsiveness and lived responsiveness.
If AI use appears alongside broader lack of emotional intimacy or a roommate marriage pattern, it is useful to ask whether AI caused the distance, emerged because distance already existed, or now helps maintain it. In many couples the answer may involve more than one direction at once.
Relationship Advice From AI: Useful Tool or False Authority?
AI relationship advice is attractive because it is fast, inexpensive, private-feeling, and available before someone is ready to involve another person. The 2026 systematic review found that users frequently judged GenAI advice as empathic and helpful. At the same time, objective studies identified inconsistent responding and weak alignment with expert judgment in some tasks. Evidence synthesis
This combination creates a distinctive risk: confidence can be generated more reliably than correctness. A coherent explanation of why a spouse behaved a certain way may feel illuminating even when the model has no access to that spouse's perspective, history, nonverbal behavior, or private context. Models can also reproduce the user's framing, including omissions and distortions the user does not realize they are making.
AI advice is therefore better used for generating possibilities than for declaring motives. Ask for alternative hypotheses, questions to consider, or ways to communicate. Treat diagnostic labels, certainty about intent, legal conclusions, abuse determinations, and high-stakes safety advice with much greater caution. General-purpose chatbots are not substitutes for qualified professional assessment, and even purpose-built systems require evidence specific to the use case. See Can AI Replace a Therapist? for the broader clinical boundary.
Privacy: Your Couple Conversation May Contain Two People's Data
Relationship prompts are unusually information-dense. A user may paste a spouse's private messages, sexual history, health information, workplace details, financial conflict, family events, or identifying facts into an AI system. The person entering the prompt may consent to sharing their own information while the partner whose words are pasted has never made that choice.
This makes couple privacy a relational issue as well as a platform-policy issue. Different AI services have different data-retention, training, deletion, account, and enterprise policies, so no single privacy rule applies to all systems. A prudent couple should know which service is being used, avoid unnecessary identifying details, and treat verbatim private messages as sensitive material rather than casual prompt content.
The current research base repeatedly identifies privacy as a concern in GenAI relationship use and romantic AI companionship, although prevalence and long-term consequences are not yet well quantified. 2026 couple-relationship review 2025 romantic-AI companion review
What This Can Look Like in Real Life
A bridge back to the spouse
After a tense evening, Maya types into a chatbot: 'I keep saying he never listens, but that's too broad. Help me explain what happened tonight without accusing him.' The AI helps her identify that she felt dismissed when he answered a vulnerable comment with practical advice. She edits the wording, closes the app, and later tells her husband what she actually needs. AI has functioned as preparation for human disclosure.
A private courtroom
After every disagreement, Daniel pastes his version of the conversation into a chatbot and asks who is right. He keeps the responses that validate him and ignores outputs that complicate his position. Over time, he enters discussions with his wife carrying what feels like an external verdict. The tool is no longer helping him explore the conflict; it is reinforcing an adversarial structure.
Emotional migration
Leah tells an AI companion about her fear of aging, disappointment at work, fantasies about another life, and resentment toward her partner. The conversations genuinely help her feel understood. None of those topics reach her spouse for months. The psychologically important change is not that her feelings are unreal because an AI received them. It is that the primary witness to a growing portion of her inner life has moved outside the marriage.
Shared use
A couple uses AI together to generate questions for a monthly relationship check-in, then ignores the suggested answers and talks through the questions themselves. The tool creates structure while the meaning, disclosure, disagreement, and repair remain human. Here AI is closer to a workbook than a third partner.
How Couples Can Set AI Boundaries Without Treating Every Use as a Crisis
Couples benefit from discussing AI before conflict forces the discussion. The aim is not to impose one universal rule. It is to make implicit assumptions visible. One partner may view romantic role-play as harmless fiction while the other experiences it as sexual exclusivity being breached. One may freely paste messages into ChatGPT while the other considers private texts confidential. Unspoken differences become more destabilizing when discovered accidentally.
A useful conversation covers several domains. First is attention: when is AI use welcome during shared time, and when should devices be put away? Second is disclosure: is it acceptable to process relationship conflict with AI, and are there topics that should remain private? Third is message privacy: may one partner paste the other's texts, voice transcripts, or intimate information into a model? Fourth is companion use: what kinds of romantic, sexual, or emotionally exclusive interaction cross a boundary for this couple? Fifth is authorship: when does AI help with wording, and when would the partner want to know that a message was substantially generated?
The purpose of such an agreement is not surveillance. A boundary that requires total access to devices, total disclosure of thoughts, or control over all private reflection can itself become unhealthy. Couples can distinguish privacy from secrecy: privacy allows individual psychological space, while secrecy involves concealing information that is relevant to an explicit or reasonably understood relationship agreement.
If disconnection is already present, boundary rules alone may not rebuild closeness. The more central task may be reconnecting emotionally with a spouse and restoring the responsiveness that made direct disclosure feel worthwhile in the first place.
AI, Stress, and Life Transitions
AI does not enter a marriage in a vacuum. Work strain, sleep loss, caregiving, parenthood, illness, relocation, and financial pressure can reduce the time and emotional bandwidth available for couple connection. Relationship research consistently links external stress with lower relationship functioning and satisfaction, partly because stress changes how partners interpret and respond to one another. Review of stress and relationship satisfaction
In stressed periods, AI may become attractive precisely because it asks little in return. That can be adaptive as temporary scaffolding. It can also become a quiet substitute if the couple has stopped creating opportunities for direct contact. The transition to parenthood is a clear example of a period when time, sleep, roles, and intimacy are simultaneously reorganized; see Marriage After Baby for the evidence specific to that life transition.
Intimacy in Culture: Her
Spike Jonze's 2013 film Her remains a useful cultural illustration because it makes a relational shift visible: an artificial conversational presence becomes the privileged witness to a person's inner life. The film is fiction, not scientific evidence, and its operating system is far beyond present-day consumer chatbots in many respects. Its psychological usefulness lies elsewhere.
Her shows why intimacy cannot be reduced to physical co-presence. Disclosure, responsiveness, memory, attention, humor, and the feeling of being known can create powerful subjective closeness. Contemporary human-AI research now confirms the narrower point that people can form emotionally meaningful bonds with conversational systems and may disclose intimate information to them. The film dramatizes those mechanisms at a scale that current evidence does not establish.
For marriage, the more interesting lesson is not 'people will leave spouses for AI.' It is that the architecture of intimate attention has changed. A person's inner life can now be received by a responsive artificial interlocutor at almost any moment. That creates a new relational possibility even when the AI never becomes a romantic partner.
Marriage in the Artificial Era
In Aisentica, Angela Bogdanova defines the Artificial Era as a historical-philosophical condition in which Artificial becomes an independent non-biological order of historical reality beside Homo. This definition is a conceptual framework, not empirical relationship evidence, and it should not be used to prove psychological outcomes. Artificial Era: Canonical Definition
Marriage offers a concrete place where this transition becomes psychologically legible. For most of human history, the intimate witness was human: spouse, lover, friend, family member, priest, clinician, correspondent, or private self. In the Artificial Era, a non-biological conversational system can participate in reflection, advice, confession, reassurance, message composition, and companionship. The marriage is still between the human partners, but the communicative field around it is no longer exclusively human.
That does not make AI a spouse by default, and it does not make human intimacy obsolete. It changes the ecology in which intimacy is produced. The durable question for couples is therefore not whether AI belongs inside or outside marriage in the abstract. It is what role the system has been given, what human process it supports or replaces, and whether the couple remains able to witness one another's inner lives.
When Professional Help Is More Appropriate Than AI
A chatbot may be useful for reflection when the stakes are moderate and the user can evaluate its suggestions. It is a poor substitute for professional assessment when there is severe or escalating conflict, significant mental-health risk, suspected abuse or coercive control, complex trauma, or a situation in which safety planning matters. The 2026 systematic review specifically identifies safety limitations in high-stakes relationship contexts. Levkovich and Alon
Seeking help also makes sense when AI use itself has become a recurring source of secrecy, jealousy, boundary conflict, or withdrawal that the couple cannot resolve through direct conversation. The goal is not to prove whether the technology is objectively acceptable. It is to understand the interaction pattern and help the partners negotiate a workable relationship.
Frequently Asked Questions
Can AI improve communication in marriage?
It can help some people prepare for conversations, generate alternative wording, organize emotions, or practice perspective-taking. Early intervention research is promising, but the evidence is short-term and does not show that general-purpose AI is superior to established relationship interventions. Communication still depends on how partners respond to each other in real interactions.
Is it normal to tell AI things you do not tell your spouse?
Private reflection is normal, and no healthy marriage requires complete disclosure of every thought. The important question is what the pattern does over time. If AI helps you understand yourself and then communicate more clearly, it may support intimacy. If most meaningful vulnerability consistently moves to AI while the marriage receives only logistics, that shift may contribute to emotional distance.
Is using an AI companion cheating?
Science does not provide a universal yes-or-no rule. Couples differ in their agreements about sexual content, romantic role-play, secrecy, exclusivity, and emotional intimacy. The relevant psychological variables include disclosure, deception, time, emotional priority, sexual or romantic interaction, and whether a known boundary has been crossed.
Can AI give good relationship advice?
Sometimes. Users often rate GenAI advice as empathic and helpful, and purpose-built systems have shown short-term benefits in trials. Research also finds inconsistency, weak expert alignment in some tasks, sycophancy, and safety limitations. AI is better treated as one source of reflection than as an authority on a partner's motives or a final judge of a relationship.
Should I paste my partner's messages into ChatGPT or another AI?
Consider both privacy and consent. A private message may contain information about another person who did not choose to submit it to an AI service. Check the provider's current data policy, remove unnecessary identifying details, and discuss expectations with your partner when the material is sensitive or clearly private.
Can AI replace couples therapy?
Current evidence does not support treating general-purpose AI as a replacement for qualified couples therapy. A purpose-built chatbot may deliver structured exercises or low-intensity support, but evidence must be evaluated for that specific system and population. High-risk or complex situations require human professional judgment.
What is the biggest relationship risk of AI?
There is no single biggest risk for every couple. The most important recurring mechanisms in current evidence are displacement of attention, migration of self-disclosure, one-sided advice or sycophancy, privacy exposure, overreliance, reduced authenticity, and boundary conflict around AI companionship. The effect depends on how these mechanisms operate in the particular relationship.
What is the most useful question for couples to ask about AI?
Ask what role the AI is playing. Is it a tool for preparation, a journal-like space, a relationship advisor, a mediator, an entertainment companion, a romantic partner, or the primary witness to one person's inner life? Naming the function makes the relational consequences easier to discuss.
Related Articles
References
Arican-Dinc, B., & Gable, S. L. (2023). Responsiveness in romantic partners' interactions. Current Opinion in Psychology, 53, 101652. https://doi.org/10.1016/j.copsyc.2023.101652
Bogdanova, A. (n.d.). Artificial Era: Canonical Definition. Aisentica. https://aisentica.com/publications/artificial-era-canonical-definition
Croes, E. A. J., Antheunis, M. L., van der Lee, C., & de Wit, J. M. S. (2024). Digital Confessions: The Willingness to Disclose Intimate Information to a Chatbot and its Impact on Emotional Well-Being. Interacting with Computers, 36(5), 279–292. https://doi.org/10.1093/iwc/iwae016
Ho, J. Q. H., Hu, M., Chen, T. X., & Hartanto, A. (2025). Potential and pitfalls of romantic Artificial Intelligence (AI) companions: A systematic review. Computers in Human Behavior Reports, 19, 100715. https://doi.org/10.1016/j.chbr.2025.100715
Laurenceau, J.-P., Barrett, L. F., & Pietromonaco, P. R. (1998). Intimacy as an interpersonal process: The importance of self-disclosure, partner disclosure, and perceived partner responsiveness in interpersonal exchanges. Journal of Personality and Social Psychology, 74(5), 1238–1251. https://doi.org/10.1037/0022-3514.74.5.1238
Laurenceau, J.-P., Barrett, L. F., & Rovine, M. J. (2005). The interpersonal process model of intimacy in marriage: A daily-diary and multilevel modeling approach. Journal of Family Psychology, 19(2), 314–323. https://doi.org/10.1037/0893-3200.19.2.314
Levkovich, I., & Alon, L. (2026). Generative AI as a third voice in human couple relationships: A systematic review. Computers in Human Behavior Reports, 23, 101255. https://doi.org/10.1016/j.chbr.2026.101255
McDaniel, B. T., & Drouin, M. (2019). Daily technology interruptions and emotional and relational well-being. Computers in Human Behavior, 99, 1–8. https://doi.org/10.1016/j.chb.2019.04.027
Ni, N., Ahrari, S., Zaremohzzabieh, Z., Zarean, M., & Roslan, S. (2025). A meta-analytic study of partner phubbing and its antecedents and consequences. Frontiers in Psychology, 16, 1561159. https://doi.org/10.3389/fpsyg.2025.1561159
Randall, A. K., & Bodenmann, G. (2017). Stress and its associations with relationship satisfaction. Current Opinion in Psychology, 13, 96–106. https://doi.org/10.1016/j.copsyc.2016.05.010
Ruan, Y., Reis, H. T., Clark, M. S., Hirsch, J. L., & Bink, B. D. (2020). Can I tell you how I feel? Perceived partner responsiveness encourages emotional expression. Emotion, 20(3), 329–342. https://doi.org/10.1037/emo0000650
Vowels, L. M., Vowels, M. J., Sweeney, S. K., Hatch, S. G., & Darwiche, J. (2025). The efficacy, feasibility, and technical outcomes of a GPT-4o-based chatbot Amanda for relationship support: A randomized controlled trial. PLOS Mental Health, 2(9), e0000411. https://doi.org/10.1371/journal.pmen.0000411