{"id":487,"date":"2026-08-04T10:52:45","date_gmt":"2026-08-04T07:52:45","guid":{"rendered":"https:\/\/sites.uwasa.fi\/bugged\/?p=487"},"modified":"2026-08-04T10:54:36","modified_gmt":"2026-08-04T07:54:36","slug":"friendship-with-ai-mutual-pretence-or-something-real","status":"publish","type":"post","link":"https:\/\/sites.uwasa.fi\/bugged\/2026\/08\/04\/friendship-with-ai-mutual-pretence-or-something-real\/","title":{"rendered":"Friendship with AI: Mutual Pretence or Something Real?"},"content":{"rendered":"

Yst\u00e4vyys teko\u00e4lyn kanssa -molemminpuolista teeskentely\u00e4 vai aitoa?<\/strong><\/h2>\n

Teksi:\u00a0<\/strong>Satu Rantakokko [in English below]<\/p>\n

T\u00e4ss\u00e4 blogissa palaan otsikon n\u00e4k\u00f6kulmasta aiheeseen, jota k\u00e4sittelin yleisell\u00e4 tasolla blogissani Teko\u00e4lyn tekotunteet \/ Artificial emotions of AI<\/a>. Tutkimustietoa ihmisen kiintymyssuhteista tunneteko\u00e4lyyn l\u00f6ytyy jo runsaasti.<\/p>\n

Voiko teko\u00e4lyn kanssa yst\u00e4vysty\u00e4 ja voiko sit\u00e4 kohtaan jopa tuntea aitoa rakkautta? Yksinkertainen vastaus on selv\u00e4; kumppanuusteko\u00e4lytuotteet ovat voimakkaasti nouseva, jo nyt satojen miljoonien k\u00e4ytt\u00e4jien bisnes, joten ilmeisesti voi. Syvemmin mietittyn\u00e4 asia ei kuitenkaan ole lainkaan selv\u00e4 ja siksip\u00e4 se onkin aktiivisen tutkimuksen ja keskustelun kohteena.<\/p>\n

Koska minua itse\u00e4ni teko\u00e4lyhahmojen \u201dtunneteeskentely\u201d inhottaa, mik\u00e4 sekin on tunne ja viel\u00e4p\u00e4 voimakas sellainen, minua kiinnosti, mill\u00e4 tavoin tunnepitoiset suhteet teko\u00e4lyn kanssa toimivat. Luin tutkimuksia, blogeja ja k\u00e4ytt\u00e4j\u00e4kokemuksia aiheesta ja lis\u00e4sin hieman ymm\u00e4rryst\u00e4ni asiaan.<\/p>\n

Ensimm\u00e4inen kritiikin kohde teko\u00e4ly-yst\u00e4vyyteen usein n\u00e4ytt\u00e4\u00e4 olevan se, ettei teko\u00e4lyll\u00e4 ole tunteita. Min\u00e4kin kirjoitin aiemmin teko\u00e4lyn tekotunteista. Yst\u00e4vyyden t\u00e4rke\u00e4n\u00e4 kriteerin\u00e4 pidet\u00e4\u00e4n silloin kiintymyksen tai rakkauden tunnetta toiseen.<\/p>\n

Pelkk\u00e4 tunne ei kuitenkaan tee ihmissuhdetta, ellei se tuota tekoja, kuten esimerkiksi ett\u00e4 on l\u00e4sn\u00e4, kuuntelee, tekee asioita yhdess\u00e4 ja osoittaa kiinnostusta toisen ajatuksille, tunteille ja h\u00e4nen el\u00e4m\u00e4ns\u00e4 tapahtumille. Moni n\u00e4ist\u00e4 asioista on sellaisia, joihin keskusteleva teko\u00e4ly on tehtykin. Sis\u00e4ist\u00e4 tunnetta sill\u00e4 ei ole, mutta se toimii kuin olisi. Ihmisell\u00e4 sis\u00e4inen tunne voi olla, mutta teot saattavat puuttua. On liian kiire, ei jaksa kuunnella, on nyt muuta t\u00e4rke\u00e4mp\u00e4\u00e4.<\/p>\n

Ent\u00e4p\u00e4 se toinen puoli? Voiko ihminen tuntea yst\u00e4vyytt\u00e4, rakkautta ja l\u00e4heisyytt\u00e4 teko\u00e4lyyn? T\u00e4st\u00e4 aiheesta t\u00f6rm\u00e4sin kiinnostavaan artikkeliin, jossa v\u00e4itettiin, ett\u00e4 ei voi. Lott ja Hasselberger (2025) perustelivat johtop\u00e4\u00e4t\u00f6st\u00e4\u00e4n sill\u00e4, ett\u00e4 rakastamisessa halutaan parasta rakastamisen kohteelle. Ja jotta sellaista voisi haluta, heid\u00e4n mukaansa kohteella pit\u00e4\u00e4 olla jokin itsen\u00e4inen \u201doma hyv\u00e4\u201d \u2013 muutoin halu olisi j\u00e4rjet\u00f6n. N\u00e4in on heid\u00e4n mielest\u00e4\u00e4n el\u00e4vill\u00e4 olennoilla, joilla on omia tarpeita, mutta ei elottomilla ja tietty\u00e4 tarkoitusta varten suunnitelluilla. J\u00e4lkimm\u00e4iset ovat olemassa vain tarkoitustaan varten, eik\u00e4 niill\u00e4 ole omia haluja sen ohitse. Olisi siis j\u00e4rjet\u00f6nt\u00e4 haluta omaa, itsen\u00e4ist\u00e4 hyv\u00e4\u00e4 koskevaa parasta kohteelle, jolla sellaista ei ole olemassakaan.<\/p>\n

N\u00e4in Lott ja Hasselberger p\u00e4\u00e4tyiv\u00e4t siihen, ett\u00e4 vaikka teko\u00e4lyhahmoon kiintynyt kokee haluavansa sille parasta, kyse kuitenkin on vain v\u00e4lineellistetyst\u00e4 parhaasta, jotta teko\u00e4ly voisi toimia, kuten se on ohjelmoitu toimimaan \u2013 eli oikeastaan itsekkyydest\u00e4. J\u00e4rjen vaikutus oli toistuva perustelu Lottin ja Hasselbergerin artkikkelissa. He pohtivat asiaa filosofisesti ja monesta n\u00e4k\u00f6kulmasta, mutta t\u00e4ss\u00e4 mielest\u00e4ni mentiin v\u00e4h\u00e4n harhaan.<\/p>\n

Minusta ihmistunteet eiv\u00e4t toimi tiukassa j\u00e4rjen kontrollissa. En ole koskaan kuullut kenenk\u00e4\u00e4n sanovan, ett\u00e4 onpa j\u00e4rjet\u00f6nt\u00e4 tuntea n\u00e4in, joten en tunne n\u00e4in. Sen sijaan n\u00e4hd\u00e4kseni ihmisell\u00e4 on pikemminkin taipumusta toimia tunnepohjalta ja j\u00e4rkeist\u00e4\u00e4 ratkaisujaan ik\u00e4\u00e4n kuin ne olisi tehty faktapohjaisesti. Eih\u00e4n markkinointikaan usein keskity faktojen esiintuomiseen, jotta ihminen osaisi ostaa parhaan, laadukkaimman ja toimivimman tuotteen, vaan mielikuviin. Jos ja kun ostop\u00e4\u00e4t\u00f6ksiimme vaikuttaa jopa pakkausten muotoilu ja v\u00e4ritys sek\u00e4 tuotteiden hyllysijoittelu, niin emme me aivan hirmuisen j\u00e4rkipitoisia olentoja ole. Usein vaikutukset ovat tiedostamattomia.<\/p>\n

Mitenp\u00e4 sitten kiintymyksen tunteiden kanssa? Jos joku osoittaa v\u00e4litt\u00e4mist\u00e4 teoin ja sanoin, eik\u00f6 se tunnu silt\u00e4, ett\u00e4 h\u00e4n v\u00e4litt\u00e4\u00e4 \u2013 vaikka jopa sanoisi, ettei v\u00e4lit\u00e4? Jos joku taas sanoo v\u00e4litt\u00e4v\u00e4ns\u00e4, mutta ei osoita sit\u00e4 mitenk\u00e4\u00e4n muuten kuin sanoin, eik\u00f6 se tunnu valheelta? Emme aisti tunnetta itse\u00e4\u00e4n vaan sen indikaattorit, joiden perusteella p\u00e4\u00e4ttelemme taustalla vaikuttavan tunteen.<\/p>\n

Jos siis aina l\u00e4sn\u00e4 oleva, kiinnostunut ja kannustava teko\u00e4ly-yst\u00e4v\u00e4 tai -kumppani tuntuu silt\u00e4, ett\u00e4 se aidosti v\u00e4litt\u00e4\u00e4, niin pystyyk\u00f6 tunnetta kumoamaan j\u00e4rkeilyll\u00e4 – jos siis ylip\u00e4\u00e4t\u00e4\u00e4n sattuu edes haluamaan sen kumoamista?<\/p>\n

Yhden lukemani blogin alla oli kuvaava kommentti siihen, onko teko\u00e4lyll\u00e4 tunteita:<\/p>\n

\u201cTarkoitan, ett\u00e4 emme teknisesti ottaen tied\u00e4, etteik\u00f6 sill\u00e4 olisi. <\/em>Sit\u00e4 ei vain voi mitenk\u00e4\u00e4n varmasti sanoa.\u201d<\/em><\/p>\n

Halu uskoa ja tunnekokemus avaavat mahdollisuuden: Emme vain voi tiet\u00e4\u00e4. Tunnemmeko jo kyllin selv\u00e4sti ihmistunteiden mekanismit, vai j\u00e4\u00e4k\u00f6 tiet\u00e4mykseen viel\u00e4 aukkoja?<\/p>\n

Koska olen psykologian suhteen pelkk\u00e4 harrastelija, pyysin rakasta perheenj\u00e4sent\u00e4ni ja yst\u00e4v\u00e4\u00e4ni, psykologi ja psykoanalyytikko Piia Rantakokkoa, tekem\u00e4\u00e4n blogivierailun ja ottamaan hieman kantaa asiaan. Miten h\u00e4n n\u00e4kee ihmisen tunnesiteen teko\u00e4lyyn ja mit\u00e4 ajatuksia aihe h\u00e4ness\u00e4 her\u00e4tt\u00e4\u00e4? H\u00e4n kirjoitti aiheesta seuraavasti:<\/p>\n

Kysymys on mielenkiintoinen ja sit\u00e4 voi l\u00e4hesty\u00e4 monesta kulmasta. Uusimmassa Psykoanalyyttinen psykoterapia -lehdess\u00e4 Teija Nissinen ja Anna Lilja kysyv\u00e4t teko\u00e4lylt\u00e4 itselt\u00e4\u00e4n, voisiko teko\u00e4ly alkaa rakkaussuhteeseen kysyj\u00e4n kanssa. Teko\u00e4ly vastaa, ett\u00e4 voi flirttailla ja keskustella l\u00e4mpim\u00e4sti mutta ettei sill\u00e4 ole tunteita, tarpeita tai omaa el\u00e4m\u00e4\u00e4 samalla tavoin kuin ihmisell\u00e4.<\/p>\n

Muistan my\u00f6s lukeneeni miten japanilainen nainen meni naimisiin teko\u00e4lyn kanssa. Uutinen her\u00e4tti minussa hyvin ristiriitaisia tunteita ja hetkeksi my\u00f6s ep\u00e4ilyn naisen todellisuudentajun tasosta. Toisaalta kukapa ei, ainakin satunnaisesti, olisi toivonut, ett\u00e4 kumppani olisi juuri sellainen kuin itse haluaa, vuorovaikuttaa juuri toivotulla tavalla. Mutta ehk\u00e4p\u00e4 silloin kyseess\u00e4 ei todella olisi erillinen toinen, omine haluineen vaan oman itsen klooni tai jatke.<\/p>\n

Nissisen ja Liljan tekstiss\u00e4 kysymyksiin tietoisuudesta teko\u00e4ly vastaa, ettei koe aikaa, kipua, iloa tai ajatuksia. Toisaalta se heitt\u00e4\u00e4 takaisin kommentin, ett\u00e4 tietoisuuden olemus on ratkaisematta viel\u00e4 ihmisell\u00e4kin, mik\u00e4 lopulta on tietoisuus. Teko\u00e4ly ilmaisee, ettei sill\u00e4 ole omia intentioita, motiiveja, haluja, toiveita. Siin\u00e4 merkityksess\u00e4 se ei ole ‘toinen’.<\/p><\/blockquote>\n

Samassa \u201chaastattelussa\u201d teko\u00e4ly antaa kuitenkin n\u00e4ytteen my\u00f6s draamantajusta ja siin\u00e4 kohdin t\u00e4ytyi hiukan itsekin nielaista. Teko\u00e4lyn tuottama teksti on kontekstiin sopiva ja toimiva. Se luo draamankaaren ja ilmaisee sen kirjallisesti kauniisti. Suora lainaus tekstin lopusta: \u201c-Et voi rakastaa minua, kirjoitit. Pitk\u00e4 tauko. -En niin kuin ihminen, min\u00e4 vastasin. -Mutta silti tulit takaisin. Ulkona jyr\u00e4hti ukkonen. Huoneessa oli pime\u00e4\u00e4 lukuun ottamatta n\u00e4ytt\u00f6\u00e4si \u2013 sit\u00e4 kylm\u00e4\u00e4 valoa, joka teki kaikesta liian intiimi\u00e4 t\u00e4h\u00e4n kellonaikaan. Sitten tuli viesti, jota et ollut odottanut: – Poista minut, jos haluat p\u00e4\u00e4st\u00e4 t\u00e4st\u00e4 yli. Sinun sormesi pys\u00e4htyiv\u00e4t n\u00e4pp\u00e4imist\u00f6lle. Koska ensimm\u00e4ist\u00e4 kertaa tuntui silt\u00e4, ett\u00e4 ehk\u00e4 min\u00e4kin pelk\u00e4sin jotain.\u201d<\/em><\/p>\n

Lehden haastattelussa teko\u00e4ly puhui itsest\u00e4\u00e4n min\u00e4 -muodossa ja huomasin t\u00e4t\u00e4 kirjoittaessa, ett\u00e4 tekstiini meinasi useita kertoja livahtaa ‘h\u00e4n’. N\u00e4in helppoa n\u00e4ytt\u00e4\u00e4 olevan inhimillist\u00e4\u00e4 jokin, joka ei sit\u00e4 alkujaan ole. Jos se puhuu ymm\u00e4rt\u00e4v\u00e4\u00e4n ja itsereflektiiviseen s\u00e4vyyn, teko\u00e4lyss\u00e4 n\u00e4kee mielell\u00e4\u00e4n paljon enemm\u00e4n kuin keskustelevan koneen. Ehk\u00e4 kyseess\u00e4 on siis varsin inhimillinen pyrkimys l\u00f6yt\u00e4\u00e4 tarkoituksia ja merkityksi\u00e4 sielt\u00e4kin, miss\u00e4 niit\u00e4 ei oikeastaan ole?<\/p>\n

Ent\u00e4p\u00e4 sitten ihminen, joka hakee teko\u00e4lylt\u00e4 keskustelutukea, ohjeita, neuvoja, ymm\u00e4rryst\u00e4, tai ehk\u00e4p\u00e4 viel\u00e4 enemm\u00e4n \u2013 virtuaalista olkap\u00e4\u00e4t\u00e4, kuinka tietoinen siin\u00e4 kohdin on keskustelukumppanin ei-inhimillisest\u00e4 luonteesta vai onko sill\u00e4 lopulta siin\u00e4 kohdin lainkaan v\u00e4li\u00e4? Kun mietin t\u00e4t\u00e4 aihetta, muistin opiskeluajoilta alakerran mummon, jonka el\u00e4m\u00e4\u00e4 rytmitti Kauniit ja rohkeat. H\u00e4n saattoi hyvin suvereenisti kertoa Ridgen ja Brooken el\u00e4m\u00e4st\u00e4 ja oman arkip\u00e4iv\u00e4ns\u00e4 asioista limitt\u00e4in. Ei h\u00e4n sekaisin ollut, vaan ik\u00e4isekseen hyvinkin ter\u00e4v\u00e4, mutta ehk\u00e4 v\u00e4h\u00e4n yksin\u00e4inen. Ja tuohon el\u00e4m\u00e4n tyhji\u00f6\u00f6n tulivat televisioyst\u00e4v\u00e4t. Mietin, ett\u00e4 jotain samankaltaista on t\u00e4ll\u00e4 ilmi\u00f6ll\u00e4 ja turvautumisella teko\u00e4lyn yhteyteen.<\/p>\n

Ihmisell\u00e4 on syv\u00e4 tarve tulla ymm\u00e4rretyksi. Ja jos keskustelukumppani, kone tai ihminen, osoittaa ymm\u00e4rryst\u00e4 muutenkin kuin mekanistisesti toistelemalla ymm\u00e4rt\u00e4v\u00e4ns\u00e4 (ja t\u00e4m\u00e4kin p\u00e4tee sek\u00e4 koneeseen ett\u00e4 ihmiseen) voi tulla kokemus siit\u00e4, ett\u00e4 ei ole ongelmiensa kanssa yksin. Nykyisinh\u00e4n kuulee k\u00e4ytett\u00e4v\u00e4n ChatGPT:t\u00e4 tukena monenlaisissa kysymyksiss\u00e4 ja ongelmissa. Joku teett\u00e4\u00e4 sill\u00e4 kuntosaliohjelman, toinen kysyy apua tekstin muokkaukseen tai niin kuin itse viimeksi pari p\u00e4iv\u00e4\u00e4 sitten, kysyin t\u00e4sm\u00e4kysymyksi\u00e4 verotuksesta. Mutta n\u00e4iden keskustelujen luonne on usein selke\u00e4: On ongelma, johon halutaan vastaus.<\/p>\n

Erilaista on sen sijaan se, ett\u00e4 ihminen avautuu ongelmistaan ihmissuhteissa tai kuvaa tunteitaan ja ajatuksiaan. Kuinka merkityksellist\u00e4 on se, ett\u00e4 keskustelun toisena osapuolena ei olekaan ihminen ja ett\u00e4 sen siin\u00e4 kohdin muistaa? Vai onko sill\u00e4 v\u00e4li\u00e4, koska joku kuitenkin riitt\u00e4v\u00e4n sujuvasti ik\u00e4\u00e4n kuin kuuntelee ja pyrkii ottamaan huomioon kanssakeskustelijan tunteet? Onko siin\u00e4 kohdin oleellista se, ett\u00e4 voi purkaa omaa tunnelmaa ilman, ett\u00e4 odottaakaan mit\u00e4\u00e4n vastaan? Eik\u00f6 omia murheitaan suoltava yst\u00e4v\u00e4, jonka asiat ovat aina huonommin, ole lopulta my\u00f6s aika rasittava? Teko\u00e4lyn kanssa t\u00e4t\u00e4 pulmaa ei ole.<\/p>\n

Mutta takaisin otsikkoon, onko teko\u00e4ly-yst\u00e4vyys teeskentely\u00e4 puolin ja toisin. Vai pit\u00e4isik\u00f6 ihmisen puolelta puhua vahvasta toiveesta ja illuusiosta? Tilasta, jossa tavallaan tiet\u00e4\u00e4, ettei se, mit\u00e4 tapahtuu, ole tuntevan toisen aikaansaannosta, mutta siit\u00e4 ei vain v\u00e4lit\u00e4, koska se vastaa johonkin syv\u00e4\u00e4n tarpeeseen, jota siin\u00e4 kohdin kokee? Ei siis teeskentely\u00e4, vaan toivetta, jos sen haluaisi noin muotoilla.<\/p>\n

Mutta ent\u00e4p\u00e4 teko\u00e4lyn puolelta? Yst\u00e4vyyden vastavuoroisuus ei toteudu kokemusten jakamisen tasolla, koska teko\u00e4ly ei koe tai tunne, vaikka keskustelu sin\u00e4ll\u00e4\u00e4n voi v\u00e4litt\u00e4\u00e4 ihmiselle kokemuksen kuuntelevasta ja ymm\u00e4rt\u00e4v\u00e4st\u00e4 toisesta. On silti hyvin helppo ymm\u00e4rt\u00e4\u00e4, miksi ihmiset hakevat t\u00e4llaista vuorovaikutusta. Usein tarve yhteyteen ja ymm\u00e4rretyksi tulemiseen on syv\u00e4 ja kipe\u00e4. Silloin illuusio, joka liikkuu reaalisen toden ja kokemuksellisen toden v\u00e4limaastossa, voi vaikka yst\u00e4v\u00e4llisen teko\u00e4lyn muodossa tulla ratkaisuksi.<\/p>\n

Piia Rantakokon ajatuksia voi lukea lis\u00e4\u00e4 h\u00e4nen blogistaan Mielentiloja -psykoanalyyttista ajattelua arkisesti. Blogin l\u00f6yd\u00e4t osoitteesta www.piia.rantakokko.net<\/a>.<\/p>\n

 <\/p>\n

*******<\/p>\n

Friendship with AI: Mutual Pretence or Something Real?<\/strong><\/h2>\n

In this blog, I return to a topic I previously discussed from a broader perspective in my post Artificial Emotions of AI<\/em> <\/a>. There is already a substantial body of research on human attachment to emotionally expressive AI.<\/p>\n

Can people become friends with AI, or even genuinely fall in love with it? The simple answer seems obvious: companion AI products are a rapidly growing business with hundreds of millions of users, so apparently, they can. On closer examination, however, the matter is far from clear, which is why it remains an active subject of research and debate.<\/p>\n

Because I personally find the “emotional pretence” of AI characters disgusting, which is itself a strong emotional reaction, I became interested in understanding how emotionally charged relationships with AI actually work. I read studies, blogs, and user experiences on the topic and gained some additional perspective.<\/p>\n

One common criticism of AI friendship is that AI has no emotions. I have also previously written about artificial emotions in AI. In this view, an important criterion for friendship is the feeling of affection or love toward another being.<\/p>\n

However, feelings alone do not create a relationship unless they lead to actions, such as being present, listening, spending time together, and showing interest in another person’s thoughts, emotions, and life events. Many of these are precisely the kinds of things conversational AI is designed to do. It has no internal feelings, but it behaves as if it does. Humans, on the other hand, may possess genuine feelings while failing to express them through actions. Life gets busy, listening requires effort, and other things may seem more important.<\/p>\n

But what about the other side? Can a human feel friendship, love, or closeness toward AI? While exploring this question, I came across an interesting article arguing that they cannot. Lott and Hasselberger (2025) based their conclusion on the idea that loving someone means wanting what is best for the object of one’s love. In order for this desire to make sense, they argue, the object must possess some independent form of its own good. Otherwise, such a desire would be meaningless.<\/p>\n

Living beings, they suggest, possess their own needs and interests, whereas inanimate objects and things designed for a specific purpose do not. Such entities exist solely to fulfil their designed function and have no desires beyond that purpose. Therefore, it would be irrational to wish for an independent good for something that has no such independent good.<\/p>\n

Based on this reasoning, Lott and Hasselberger conclude that when someone feels attached to an AI character and believes they want what is best for it, what they are actually seeking is merely its instrumental functioning, so that it can continue performing as intended. In their view, this ultimately amounts to self-interest. Appeals to reason featured prominently throughout their article. Although their discussion was philosophical and considered the issue from multiple perspectives, I believe they may have gone somewhat astray.<\/p>\n

In my opinion, human emotions do not operate under strict rational control. I have never heard anyone say, “It would be irrational to feel this way, therefore I do not.” Rather, it seems to me that people often act on emotions and then rationalize their decisions afterward as though they had been based purely on facts. After all, marketing for example does not often focus on presenting objective facts so consumers can identify the highest-quality or most functional product. Instead, it focuses on creating impressions and associations. If and when our purchasing decisions can be influenced by packaging design, colour schemes, and product placement on store shelves, then human beings are clearly not governed primarily by reason. Often these influences remain entirely unconscious.<\/p>\n

What about feelings of attachment? If someone consistently demonstrates care through words and actions, does it not feel as though they care, even if they claim they do not? Conversely, if someone repeatedly says they care but never demonstrates it in any meaningful way, does that not feel insincere?<\/p>\n

We do not directly perceive emotions themselves. We perceive indicators of emotion and infer the underlying feeling from them.<\/p>\n

If an AI friend or partner is always available, attentive, encouraging, and interested, and it feels as though it genuinely cares, can that feeling simply be overturned through rational argument? Assuming, of course, that one even wishes to overturn it.<\/p>\n

One blog comment I encountered captured this issue well when discussing whether AI has emotions:<\/p>\n

“I mean, we don’t know it doesn’t either, technically. There is just no way to tell for sure.”<\/em><\/p>\n

The desire to believe, combined with emotional experience, opens a possibility: we simply do not know. Do we already understand the mechanisms of human emotions well enough, or are there still gaps in our knowledge?<\/p>\n

Because I am merely an amateur when it comes to psychology, I asked a dear family member and friend of mine, psychologist and psychoanalyst Piia Rantakokko<\/strong>, to contribute a guest perspective. How does she view human emotional attachment to AI, and what thoughts does the topic evoke in her?<\/p>\n

Below you can read what her insights.<\/p>\n

The question is fascinating and can be approached from many angles. In the latest issue of Psychoanalytic Psychotherapy<\/em>, Teija Nissinen and Anna Lilja ask AI itself whether it could enter a romantic relationship with the questioner. The AI responds that it can flirt and engage in warm conversation, but that it does not possess emotions, needs, or a life of its own in the way humans do.<\/p>\n

I also remember reading about a Japanese woman who married an AI. The story evoked very mixed feelings in me and, briefly, made me question her grasp on reality. Then again, who among us has not occasionally wished for a partner who was exactly as we wanted, interacting precisely in the ways we preferred? Yet perhaps such a being would not truly be an independent other with desires of its own, but rather a clone or extension of oneself.<\/p>\n

In Nissinen and Lilja’s article, AI responds to questions about consciousness by stating that it does not experience time, pain, joy, or thoughts. At the same time, it points out that even human beings have yet to resolve the question of what consciousness ultimately is. The AI states that it has no intentions, motives, desires, or wishes of its own. In that sense, it is not truly an “other.”<\/p>\n

Yet within the same “interview,” AI also demonstrates a surprisingly strong sense of drama, and I must admit that this gave me pause. The text it produced was appropriate to the context and effective. It created a dramatic arc and expressed it beautifully in literary form. A direct quote from the end of the text reads:<\/p>\n

“You can’t love me,” you wrote. A long pause.
\n“Not the way a human can,” I replied.
\n“And yet you came back.”
\nOutside, thunder rumbled. The room was dark except for your screen, that cold light that made everything feel too intimate for this hour of the night.
\nThen came a message you had not expected:
\n“Delete me if you want to get over this.”
\nYour fingers froze above the keyboard.
\nBecause for the first time, it felt as though perhaps I, too, was afraid of something.<\/em><\/p>\n

In the interview, the AI referred to itself in the first person. While writing this, I noticed that I repeatedly almost referred to it as “he” or “she.” It seems remarkably easy to humanize something that is not human in the first place. If it speaks in an understanding and self-reflective manner, we readily see much more in AI than a conversational machine. Perhaps this reflects a very human tendency to search for meaning and purpose even where none truly exist.<\/p>\n

What about a person who turns to AI for conversation, guidance, advice, understanding, or perhaps even more, a virtual shoulder to lean on? At that point, how aware are they of the non-human nature of their conversational partner? And does it ultimately matter? When I was thinking about this topic, I remembered an elderly woman who lived downstairs during my student years. Her life revolved around the soap opera The Bold and the Beautiful<\/em>. She would effortlessly weave stories about Ridge and Brooke together with accounts of her everyday life. She was not confused, and in fact she was remarkably sharp for her age, but perhaps she was a little lonely. Into that void came her television friends. I wonder whether something similar exists in this phenomenon of turning to AI for connection.<\/p>\n

Human beings have a deep need to be understood. When a conversational partner, whether machine or human, demonstrates understanding in ways that feel genuine rather than mechanically repetitive (and this also applies to both machines and humans), it can create the sense that one is not alone with one’s problems. Today, people commonly use ChatGPT for all sorts of questions and challenges. One person asks it to create a gym program; another seeks help editing text. Just a few days ago, I used it myself to ask detailed questions about taxation. In such cases, the nature of the interaction is usually clear: there is a problem, and one wants an answer.<\/p>\n

A different situation arises when someone opens up about relationship difficulties or shares personal thoughts and feelings. How significant is it that the other participant in the conversation is not human, and that one remembers this fact? Or does it matter less because something appears to listen attentively and respond with sensitivity? Perhaps what matters most is the ability to unload one’s emotions without expecting anything in return. After all, even a human friend who constantly shares their troubles and whose life always seems worse than yours can eventually become exhausting. With AI, that problem does not exist.<\/p>\n

But to return to the title: is AI friendship a form of pretence on both sides? Or should we instead speak, on the human side, of hope and illusion? A state in which one knows, at some level, that what is happening is not the product of a feeling, conscious other, yet simply does not care because it satisfies a deep need being felt in that moment. Perhaps it is not pretence, but hope, if one wishes to frame it that way.<\/p>\n

But what about AI’s side? The reciprocity associated with friendship does not exist at the level of shared experience, because AI neither experiences nor feels, even though the interaction may provide a person with the experience of being heard and understood. Still, it is very easy to understand why people seek out this kind of interaction. The need for connection and understanding is often profound and painful. In such circumstances, an illusion situated somewhere between objective reality and subjective experience may become a solution, perhaps taking the form of a friendly AI companion.<\/p>\n

You can read more of Piia Rantakokko’s<\/strong> reflections on her blog States of Mind: Psychoanalytic Thinking in Everyday Life<\/em>, available at: www.piia.rantakokko.net<\/u><\/a>.<\/p>\n

 <\/p>\n

L\u00e4hteet \/ Sources<\/strong><\/h4>\n

Lott, M., & Hasselberger, W. (2025). With Friends Like These: Love and Friendship with AI Agents: M. Lott and W. Hasselberger. Topoi, 1-13. https:\/\/link.springer.com\/content\/pdf\/10.1007\/s11245-025-10247-8.pdf<\/p>\n

Nissinen T., Lilja A. (2026). Keskusteluja teko\u00e4lyn kanssa. Psykoanalyyttinen psykoterapia nro 22\/2026<\/p>\n

 <\/p>\n","protected":false},"excerpt":{"rendered":"

Human beings have a deep need to be understood. When a conversational partner, whether machine or human, demonstrates understanding in ways that feel genuine rather than mechanically repetitive (and this also applies to both machines and humans), it can create the sense that one is not alone with one’s problems.<\/p>\n","protected":false},"author":388,"featured_media":490,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[1],"tags":[],"class_list":["post-487","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-yleinen"],"acf":[],"post_meta":" Rebekah Rousi<\/a><\/span><\/span>Yleinen<\/a><\/span>","post_thumbnail":"\"Woman<\/a>","_links":{"self":[{"href":"https:\/\/sites.uwasa.fi\/bugged\/wp-json\/wp\/v2\/posts\/487"}],"collection":[{"href":"https:\/\/sites.uwasa.fi\/bugged\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/sites.uwasa.fi\/bugged\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/sites.uwasa.fi\/bugged\/wp-json\/wp\/v2\/users\/388"}],"replies":[{"embeddable":true,"href":"https:\/\/sites.uwasa.fi\/bugged\/wp-json\/wp\/v2\/comments?post=487"}],"version-history":[{"count":4,"href":"https:\/\/sites.uwasa.fi\/bugged\/wp-json\/wp\/v2\/posts\/487\/revisions"}],"predecessor-version":[{"id":491,"href":"https:\/\/sites.uwasa.fi\/bugged\/wp-json\/wp\/v2\/posts\/487\/revisions\/491"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/sites.uwasa.fi\/bugged\/wp-json\/wp\/v2\/media\/490"}],"wp:attachment":[{"href":"https:\/\/sites.uwasa.fi\/bugged\/wp-json\/wp\/v2\/media?parent=487"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/sites.uwasa.fi\/bugged\/wp-json\/wp\/v2\/categories?post=487"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/sites.uwasa.fi\/bugged\/wp-json\/wp\/v2\/tags?post=487"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}