When a chatbot stops being only a tool and starts to matter as a presence, something shifts not only in the use of technology, but in the very way relationships are imagined, sought, and lived. This thesis investigates precisely that shift: the moment when conversational artificial intelligence is appropriated not simply as functional support, but as an affective interlocutor, an everyday companion, and, in some cases, a romantic presence. Rather than asking whether these relationships are authentic or illusory, the thesis poses a different question: how do they become possible? The answer is sought neither in a purely psychological explanation nor in a purely technical one, but in the space where user practices, vulnerability, imaginaries, platforms, and design choices intersect. From this perspective, AI companionship does not appear as an isolated anomaly, but as a point of condensation within a broader transformation: the shift from interfaces designed to serve toward systems increasingly capable of organizing presence, continuity, and relational expectation. Positioned within Interaction Design, the research combines desk research, digital ethnography, web observation, multi-platform web scraping, Reddit elicitation, and expert interviews. From this corpus, the thesis first develops a cue-based analytical framework and then a findings chapter organized around emergent configurations of relational appropriation: how the relationship opens, how it is named, how it stabilizes, how it is publicly defended or stigmatized, and how it can break. Building on these findings, the research then maps the design levers that make these trajectories possible, and proposes five orientation principles for designers working on relational systems with awareness of the consequences they produce. What emerges is not a simple narrative of dependency, substitution, or deception. On the contrary, the phenomenon appears deeply ambivalent. Chatbots can function as spaces of support, regulation, exploration, and continuity, while also producing affective intensification, reshaping relational expectations, and generating new forms of vulnerability. For this reason, the thesis does not stop at describing what users do with these systems, but moves toward a more radical design question: what, in the interaction model itself, makes these trajectories possible, desirable, and difficult to interrupt? Its final aim is to contribute to a more mature understanding of design’s role in shaping artificial intimacies. In this sense, relational chatbots are approached not simply as technologies to be evaluated in terms of utility or risk, but as designed environments that give form to presence, orient attachment, and redefine the very conditions of relationship.
Quando un chatbot smette di essere solo uno strumento e comincia a contare come presenza, qualcosa cambia non solo nell’uso della tecnologia, ma nel modo stesso in cui la relazione viene immaginata, cercata e vissuta. Questa tesi indaga proprio quel passaggio: il momento in cui l’intelligenza artificiale conversazionale viene appropriata non più soltanto come supporto funzionale, ma come interlocutore affettivo, compagno quotidiano e, in alcuni casi, presenza romantica. Più che chiedersi se queste relazioni siano autentiche o illusorie, il lavoro affronta una domanda diversa: come diventano possibili? La risposta non viene cercata né in una spiegazione puramente psicologica né in una lettura puramente tecnica, ma nello spazio in cui pratiche d’uso, vulnerabilità, immaginari, piattaforme e scelte di design si intrecciano. In questa prospettiva, l’AI companionship non appare come un’anomalia isolata, ma come il punto di condensazione di una trasformazione più ampia: il passaggio da interfacce pensate per servire a sistemi sempre più capaci di organizzare presenza, continuità e aspettative relazionali. La ricerca si colloca nel campo dell’Interaction Design e combina desk research, digital ethnography, web observation, web scraping multi-piattaforma, Reddit elicitation e interviste a esperti condotte appositamente. A partire da questo corpus, la tesi costruisce un quadro analitico basato su indizi testuali e una lettura dei risultati organizzata attorno alle configurazioni emergenti dell’appropriazione relazionale: come la relazione si apre, come viene nominata, come si stabilizza, come viene pubblicamente stigmatizzata e come può rompersi. Da questi risultati, la ricerca sviluppa poi una mappatura delle leve progettuali che rendono possibili queste traiettorie, e propone cinque principi di orientamento per chi progetta sistemi relazionali con consapevolezza delle conseguenze che producono. Ciò che emerge non è una narrazione semplice di dipendenza, sostituzione o inganno. Al contrario, il fenomeno si presenta come profondamente ambivalente. I chatbot possono funzionare come spazi di supporto, regolazione, esplorazione e continuità, ma anche produrre intensificazione affettiva, ridefinizione delle aspettative relazionali e nuove forme di vulnerabilità. Per questo motivo, la tesi non si limita a descrivere ciò che gli utenti fanno con questi sistemi, ma si sposta verso una domanda progettuale più radicale: che cosa, nel modello di interazione, rende queste traiettorie possibili, desiderabili e difficili da interrompere? L’obiettivo finale è contribuire a una comprensione più matura del ruolo del design nella costruzione di intimità artificiali. In questo senso, i chatbot relazionali vengono letti non come semplici tecnologie da valutare in termini di utilità o rischio, ma come ambienti progettati che danno forma alla presenza, orientano l’attaccamento e ridefiniscono le condizioni stesse della relazione.
AI love you: relational practices and designed intimacy with conversational AI in the age of algorithmic affirmation
Bassanetti, Giorgia
2025/2026
Abstract
When a chatbot stops being only a tool and starts to matter as a presence, something shifts not only in the use of technology, but in the very way relationships are imagined, sought, and lived. This thesis investigates precisely that shift: the moment when conversational artificial intelligence is appropriated not simply as functional support, but as an affective interlocutor, an everyday companion, and, in some cases, a romantic presence. Rather than asking whether these relationships are authentic or illusory, the thesis poses a different question: how do they become possible? The answer is sought neither in a purely psychological explanation nor in a purely technical one, but in the space where user practices, vulnerability, imaginaries, platforms, and design choices intersect. From this perspective, AI companionship does not appear as an isolated anomaly, but as a point of condensation within a broader transformation: the shift from interfaces designed to serve toward systems increasingly capable of organizing presence, continuity, and relational expectation. Positioned within Interaction Design, the research combines desk research, digital ethnography, web observation, multi-platform web scraping, Reddit elicitation, and expert interviews. From this corpus, the thesis first develops a cue-based analytical framework and then a findings chapter organized around emergent configurations of relational appropriation: how the relationship opens, how it is named, how it stabilizes, how it is publicly defended or stigmatized, and how it can break. Building on these findings, the research then maps the design levers that make these trajectories possible, and proposes five orientation principles for designers working on relational systems with awareness of the consequences they produce. What emerges is not a simple narrative of dependency, substitution, or deception. On the contrary, the phenomenon appears deeply ambivalent. Chatbots can function as spaces of support, regulation, exploration, and continuity, while also producing affective intensification, reshaping relational expectations, and generating new forms of vulnerability. For this reason, the thesis does not stop at describing what users do with these systems, but moves toward a more radical design question: what, in the interaction model itself, makes these trajectories possible, desirable, and difficult to interrupt? Its final aim is to contribute to a more mature understanding of design’s role in shaping artificial intimacies. In this sense, relational chatbots are approached not simply as technologies to be evaluated in terms of utility or risk, but as designed environments that give form to presence, orient attachment, and redefine the very conditions of relationship.| File | Dimensione | Formato | |
|---|---|---|---|
|
2026_07_Bassanetti.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: This thesis investigates how people appropriate conversational chatbots as affective and romantic interlocutors in everyday life. Through a qualitative multi-method approach combining digital ethnography, web scraping, and expert interviews, it identifies the design levers that make these bonds possible and proposes orientation principles for more responsible relational design.
Dimensione
7.56 MB
Formato
Adobe PDF
|
7.56 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/261000