"Ukuhlukaniswa, ukujabulela-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo.
Byung-Chul Han, I-Burnout Society
"Ukuhlukaniswa, ukujabulela-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo-umthamo.
U-Byung-Chul HanI-Burnout Society
Ngingathanda embhedeni ku-1:43 amahora, umzimba wakho usebenza okuhle lapho kubhalwe ku-playlist engatholakalayo yintoni ukuthi ungayifuna.
Uyakwazi ukuxhumana nabathengi, kodwa ungacindezela. Uyavula i-portail ye-therapist yakho, kodwa i-slot yokuqala iyisikhashana elilandelayo. Uyakwazi ukuxhumana, kodwa i-notebook ayikwazanga yini.
Ngakho-ke, ngexesha elihlukile, elihlukile-Mary, ufake i-ChatGPT.
“Ukuhlukanisa yini kungenzeka nangona kunjalo?” ungathanda.
"Ngingathanda ukuthi kungcono ukuthi kungcono," ungathanda.
"Ngingathanda ukuthi kungcono ukuthi kungcono," ungathanda.
Ngiyazi. Ukuphendula akuyona, noma ngisho ngokucacileyo. Kodwa kuyinto enhle. Enhle. Reflective. Kuyinto engcono kunazo. Ngcono kunazo. Ngcono kunazo. Ngaphezulu kunazo. Ngaphezulu kunazo. Ngaphezulu kunazo.
Uyakwazi ukuxhumana. Uyakwazi ukuxhumana. Ngaphezulu kunazo mayelana nawe, kuncike kahle.
Umehluko, okungenani emangalisayo, igxile emhlabeni wonke. Futhi akuyona kuphela abantu awuthanda - akuyona kuphela ama-tech bros noma ama-weirdos emzimbeni - kodwa wonke umuntu owayenza ukuthi wahlukaniswa emahoreni emangalisayo, okwenziwe kakhulu ukujula, okungenani kakhulu ukujula, futhi okwenziwe kakhulu ukwenza noma yini, kodwa ukhompyutha ibhokisi elifanele ukuthi kungenzeka.
Kuyinto lokuphelelwa kwezinto ezihambayo ezivela ku-inthanethi ye-AI ngenxa yokumelana. Akukho ngenxa yokuzonwabisa. Akukho ngenxa yokuzonwabisa. Kodwa ngenxa yokuba, ehlabathini enhle futhi enhle kakhulu, kukhona.
Thina siphila isikhathi eside yokubambisana nezempilo emzimbeni, futhi isikhathi eside yokufumana ukuncintisana. Ukwelashwa kunezingeni. Amadivayisi zihlanganisa. I-economy yakhula. I-brain yakho yakhula. Futhi ngakho-ke lapho umuntu wathi, "Ukuhlola, iyahambisana engcono kunomzimba yami," akuyona kuphela umxokozelo. Kuyinto i-simptom.
Yini usebenza kanjani? Amasihlobiso zemvelo zihlabathi ukufakelwa ngokushesha?
1. Thola i-shrink yakho entsha: i-chatbot
You're not alone if you've ever found yourself trusting to these bots.
I-AI, ikakhulukazi izixhobo ze-ChatGPT, isetshenziselwa abantu, njengomthombo we-emotional. Akukho kuphela ukulandwa kwe-imeyili noma ukuphefumula ama-notes ye-meetings.UkuhlobisaUkusekela. Ukuze ama-breakdowns emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni. It is not science fiction – it is a Reddit thread, a YouTube confession, a tweet that reads:
Futhi isampula iyatholakala. Abantu abasetyenziselwa ukwesekwa kwesifiso? I-positive kakhulu. Zihlanganisa ukuthi akuyona-non-judgmental, ngokuqhubekayo, futhi enhle ngokuzinzakalelayo. Ngaphezu kwalokho, ama-sceptics - ama-sceptics abanamakhasimende abasetyenziswa - zihlanganisa umqondo. "It is just a robot." "It is not.UkucaciswaUkusiza. “Ukuhlezi okuhle.”
Kodwa lapha ingxenye emangalisayo: Amabhodlela amabhodlela ngokufanele.
Thola thwebula ama-characters ku-thera-adjacent ecosystem entsha:
a. Umphumela we-evangelist
Kuyinto umntu owenziwe ngempumelelo kumandla we-AI for emotional support. Okokuqala engabonakali, wahlanganyela ku-ChatGPT ngesikhathi kwemizuzu ye-stress futhi wahlanganyela into emangalisayo: ibhalisele, futhi ibhalisele kakuhle. Ukuze kwabo, ukubhuka kubhalwe emzimbeni we-non-judgmental, ubumnandi we-bot - ephelele kumahora emzimbeni ebusuku emzimbeni lapho zihlola futhi ufuna umuntu ukuxhumana.
I-AI Support Evangelist iyinhlangano yokuxhumana izipho zabo, ukunakekela i-chatbot kumadokhumenti owayenza. Bafuna ukuthi i-AI inokukwazi ukunikezela okuhle ama-connections embalwa: ukunakekelwa okuqhubekayo, okuhlobene. Ngokuthi, bayazi ukuthi akuyona ingxenye ye-therapy, kodwa iyinhlangano enhle lapho ukuxhumana kwamanani ayikho.
Zithi AI kuyinto ukwelashwa-ke konke, kodwa kubo, kuyinto engcono kunoma. It is always available, never critical, and free from the inconvenience that sometimes comes with opening up to another person.
B. Umthandazo we-Skeptic
I-Skeptic has a difficult time accepting the idea of a chatbot providing emotional support. Thina ukunakekela izinzuzo ku-AI, kodwa lapho kufinyeleleke into enhle njengokungcebeleka kwesifo se-mental, baye zihlanganisa. Zihlanganisa ukuthi ngaphandle kokuphumelela okwenziwe futhi iziqinisekiso ze-therapist enhle, i-AI iyatholakala emibi yayo yokufunda futhi ngokwenene ukunceda ngezimo ezinzima.
I-Skeptic isibuyekeza ukujabulela umntu kanye nokuxhumana, futhi ibonise ukuthi umatshini, kungekho okuqhubekayo, ayikwazi ukunikezela imibuzo noma ukujabulela i-empathy enobuchwepheshe, noma ngisho umbhali owaziwa.
Ngokuvamile, ukweseka kwe-AI ye-emotional iyinhlangano, okungenani, kodwa akuyona isakhiwo esiyingqayizivele yokuxhumana kwamandla.
c. Umdlali we-casual
Uhlobo lwe-inthanethi ye-inthanethi ye-inthanethi ye-inthanethi ye-inthanethi ye-inthanethi ye-inthanethi ye-inthanethi ye-inthanethi ye-inthanethi ye-inthanethi ye-inthanethi ye-inthanethi ye-inthanethi.
I-User ye-Casual ayikwazanga izixazululo zokuguqulwa kwezilwane; bafuna kuphela isivumelwano esincane noma mhlawumbe isivumelwano esizayo esiza ku-stress.
Thina akubona njengesithuthuthu, futhi akufuna ukucindezeleka okuhlobene. Kuyinto nje isixhobo olufanelekayo lapho akufuneka ukubonisa imibuzo yabo ngaphandle kokuqinisekisa. Akukho ukuxhaswa emoyeni, nje ukuxhaswa okusheshayo ukucindezelisa umqondo yabo.
Ngezinye, lezi ezintathu abantu zihlanganisa umkhosi wokuqala we-therapy-adjacent ecosystem-one ayihlaziywa ngempumelelo olungazwe ku-AI yokwelapha umthamo, kodwa nge-patchwork ye-negotiations yomsindo, ukufinyelela, nokuphepha.
Waze akufanele ku-AIUkulungeleI-one sees a lifeline, another a gimmick, and the third, just a tool to get through the day.
Kodwa ngenxa yayo zonke izidakamizwa, kukho into eyodwa akakwazi ukunceda: isisindo esisodwa. Ukulungele, ukujabulela, ngokushesha ngokushesha ku-empathy.
Lezi ama-persona angakwazi ukujongana mayelana nemali yayo, ama-limits yayo, ngisho ne-ethics yayo - kodwa bonke bafikelele ngokufanele into efanayo:
Kuyinto okuhlobene abantu.
Futhi lokhu - ngaphezulu kwezinye izinto - kuyinto enza kakhulu. Futhi kahle kakhulu.
2. I-The Great Pretender
"Ukuba i-ChatGPT namanye ama-AI amamodeli akufanele ukudluliselwa okuyi-humane?" iyiphi imibuzo eyenziwe ngokuvamile.
Kodwa sicela ukuguqulwa kwalo ukusuka kwelinye inkinobho yayo, lapho akukho nto ekuqaleni:
"Ukuza kanjani into enhle ngokucacisa futhi ngokucophelela njenge umntu?"
Njengoba nezinye izimo zokusebenza ezininzi, imibuzo enhle kakhulu futhi ngokuvamile sikhuseleko ngokuvamile kubaluleke emzimbeni. Lokhu kwenziwe njalo. Sihlola izindandatho ngaphambi kokwakha izindiza. Sihlola kanjani izindandatho ezikhuthaza ngaphambi kokucwaninga i-robots ezithile.
Futhi lapho kufinyelelwa ne-language ne-think, inguqulo enhle kakhulu futhi cishe njalo kulungile kuyinto ukuchithwa kwe-brain yabantu.
Kuyinto ukuthi abacwaningi bafuna ukwenza - ukwakha izinjini ezibonisa ukuthi sincoma, ukhangela, futhi ukufundisa. Akungekho ngokuphelele, akungekho ngokuphelele. Kodwa okungenani kakhulu ukuthi ngamanye amaxesha unamathele ukuthi akuyona abantu.
A. Umzimba we-digital
I-Artificial Neural Networks (ANNs) iyinhlangano yethu enhle yokufaka indlela yokusebenza kwe-brain yabantu, ngaphandle kokusebenza ngokwenene njenge-brain.
Yenziwe emzimbeni yakho njengomgca we-neurons eyenza ama-scars electrical phezu ama-trillions ye-connections. Uma uzobona umculo, ukhangela umculo, noma ungenza umdlavuza, lezi ama-neurons zenza umdlavuza we-synchronized eyenziwe ngexabiso.
I-ANN isitholela le blueprint – uhlobo. Ngaphandle kwe-neurons, zisebenzisa ama-nodes. Ngaphandle kwe-dendrites ne-axons, zisebenzisa ama-weights kanye ne-biases. Ngaphandle kwalokho, i-metaphor yakhelwe: isitholela idatha, isitholele ngezigaba eziningana (noma ezingu-hundreds), ukuguqulwa kwezingane, ukucindezeleka, ukuguqulwa. Lokhu kubizwa ngokuthiUkuqeqeshwa.
Kodwa lokhu akuyona kuphela amanethiwekhi. Amasipala ezinkulu – njenge-ngaphandle kwe-ChatGPT – kuyinto enhle. Sihlola izigidi ze-parameters. Kuyinto efana nokwakha umlilo we-mechanical kanye nokufunda ngempumelelo ukuthi uyenze i-cyborg dragon.
I-Oppenheimer ye-AI
Kwangathi ngo-2017, iqela le-Google ingcindezi wabhala iphepha nge-inducing-sleep“Ukuhlobisa kuyinto konke kufuneka.”Ukukhetha okungenani kakhulu ngexesha elandelayo - ngokuvamile akuyona uhlobo leyo ungathanda ukujabulela izakhiwo yokufakelwa kwegama. Ukuhlola kulokhu kubona Oppenheimer wokuqala wathwala i-atomic bomb yayo.
Kodwa ngaphansi kwegama wamafutha wamafutha wamafutha wamafutha wamafutha: i-Transformer.
I-Transformer Architecture yasebandakanya konke. Ngaphezu kwalokho, ama-networks ze-neural zibonisa ulimi njenge-children eyenza i-novel – zihlanganisa izinhlamvu eyodwa ngenyanga futhi zihlole ngokushesha.
Kwakungabikho imemori, umklamo, kanye ne-nuance. Ngithole isampula se-old school,"Ndiyazi ukuthi indoda yami ivame. Ngiyazi, ngihambe ku-park,"futhi angakwazi ukuguqulwa nge“I-Park kuyinto enhle!”—Ukuhlukanisa okungaziwa kwelanga yakho emoyeni.
I-Transformers yasungulwa umkhakha ebizwa ngokuthi i-"Attention", okuvumela amamodeli ukuhlola imibuzo phakathi kwezindaba kanye nokufunda umklamo njengoba abantu.
Ngaphezu kwalokho, model ingaba wathi,"Ukuhlukanisa ukufunda lokhu. Ngamanye izikhathi ukujabulela kungasiza."Ngemuva kwalokho, kubonakala ukuthiThola itNgemuva kwalokho, kuyinto engaphansi njengama-autocomplete futhi engaphezulu njengama-friend enikezela.
c. I-hydrogen bomb
Futhi uma thwebula mayelana ne-Transformers njengoko i-bomb atomic ye-AI, ke i-ChatGPT iyona i-Hydrogen Bomb. Yakhelwe ku-foundation futhi enhle kakhulu.
I-ChatGPT iyinhlangano olulodwa ngqo le-Transformer revolution. It kuyinto ebizwa ngokuthi i-Large Language Model, noma i-LLM. Yonke umsebenzi yayo kuyinto ukuhlaziywa igama elilandelayo ngegama elandelayo, ngokuvumelana nezinto ezaziwa. Akukho umzimba, akukho umqondo - nje umshini wokubacindezelisa imibhalo enhle kakhulu.
Kodwa ukwandisa - ukunakekela amabhuku, ama-scripts ezimbonini, ama-transcripts ye-therapy, ama-tragedies e-Shakespeare, ne-flame wars kusuka ku-Reddit - futhi into emangalisayo ikhona. Kuyashintshele ukhangela njenge-machine. Kuyashintshele njenge-humane.
Kuyinto ukuthi akuyona ukuthiWazenoma yini, kodwa ngenxa yokuba i-language efanelekayo ukubonisa umdlalo ngokuvumelana.
I-Researchers ibizwa ngokuba lo mphumela emangalisayo - izinzuzo ezingenalutho ezivela emakhemikhali emikhulu, njenge-chats, empathy, kanye nezinhlangano ezinzima. Akukho ngenxa yokuxhumana kwakho, kodwa ngenxa yokuxhumana ngokuvamile.NgiyaMhlawumbe ufuna ukufunda.
Thina ngokweneneNgena ngemvumeYini?
I-John Searle's i-Chinese Room i-argument ibonise umfanekiso: ukubonisa ukuthi ungathandwa emkhumbi, ukunikezela ama-character Chinese nge-slot. Uyakwazi ukuxhumana ne-language, kodwa unayo i-rulesbook enikezela indlela yokusabela ngamunye "in" futhi ukunikezela "out" ngokuvamile.
Ukuze ama-outsiders, kubonakala ukuthi uxhumane. Ngezinye, wena kuphela shuffling ama-symbols.
Kuyinto i-ChatGPT. I-illusion ye-understanding ngaphandle kwe-awareness.
Nokho, uma ushiye ukuthi ungacindezeleka, inikezela ngezindlela ezinzima. Ngenxa yokufunda izibonelo ezigidi zokuzonwabisa. It uyazi indlela yokuzonwabisa. I-cadence ye-heartbreak. I-rituals ye-linguistic etholakalayo lapho siphindezeleka.
I-Sherry Turkle iyiprofesa e-MIT kanye ne-psychologist owenziwe iminyaka eminyakeni ukubuyekeza indlela abantu abahlanganisa ubuchwepheshe. I-She once wrote that sometimes, we don't need someone who truly understands — we just want someone who pretends to.
Futhi i-ChatGPT, okungenani okungcono noma okungcono, kuyinto enhle ukuhlangabezana.
3. Yini iveza abantu abaningi
I-Anthropomorphizing yonke into
Ngiyazi ukuthi kuyinto real? Yini umshini angakwazi ukujabulela kakhulu umntu?
Ngokuba thina kubhalwe ngokufanelekileyo, sinikeza konke.
Thina bafumene ama-Roombas ethu. Thina thuthile ama-robots emangalisayo. Thina izibopho emangalisayo emangalisayo emizini we-Pixar. Thina umatshini owaziwa ngokuthi nathi, futhi thina siphinde ukuthi kukhona 'omunye' emangalisayo emangalisayo.
ikhaya
ikhaya
Kuyinto ingxenye aImininingwaneI-terme eyasetyenziselwa ukubonisa amafane abalandela ngenxa ye-celebrity isixazululo isixazululo se-chatbot, njengoba abasebenzisi abalandela amangqamuzana zabo futhi bafuna imibuzo malunga ne-meaning of life.
I-Machine ayazi, kodwa ayazi ukuthi umntu angazange wathi uma wahlala.
Kodwa, lokhu kuyinto enhle futhi konke, akugcwalisa ngokugcwele ukuthi yini bonke lezi abantu zihlangene ku-instrument eyenziwe ukuthi awukwenzelwe ukuba yi-therapist?
I-Outsourcing ye-Emotional Labour ku-Machines
Ngenxa yokuba thina manje ngempumelelo futhi lonely. Ngaphezu kwalokho kunoma.
Ngokusho imicimbi ye-AI ku-mental well-being, le chart ibonisaYiniUma ama-adult 4 kusuka ku-5 ye-Generation Z akufanele ukunemba, akugqiba akufanele izindlela ezintsha zokuxhumana - ngisho kusukela nge-chatbot.
Ukukhishwa kwe-Support Evangelist, i-Dubtful Skeptic, ne-Casual User, njengokungapheli, kubaluleke kakhulu ngokusebenzisa le lens: lapho ukunambitheka kubaluleke kakhulu, ngisho kwezinto ezingenalutho ezifana ne-AI zihlanganisa njenge-liveline.
Thina ngempumelelo, ngempumelelo, futhi ngempumelelo.
Futhi ukwelashwa - uhlobo efanelekayo - kuyinto enhle, enzima, futhi ngokuvamile engatholakali. Izindaba eziningi zihlanganisa ukuthi ukuzinza kwezimali kanye ne-cost of living ye-crisis zihlanganisa abantu abadala ukufinyelela nokuxhumana nezidingo, kuhlanganise izinsizakalo zempilo
Ngexesha elandelayo, ukweseka kwezingane zihlanganisa izikhwama zayo zayo. Wonke abantu zihlanganisa. Wonke abantu akufanele ukujabulela.
Futhi i-chatbot iyisisombululo ephelele. Zonke isikhathi. Zonke isikhathi "ukucindezela." Engapheliyo, engapheliyo, engapheliyo engapheliyo ukuba ufinyelele ukuze inokufunda into.
Ngaphezu kwalokho, kulungile.
Futhi sociologically ukhangela, sinamathela isixazululo seismic. I-labour emoyeni - umsebenzi enhle, encibilikayo yokufunda, ukuvalwa, nokugcina indawo - iyathunyelwa kumakhasimende.
Uma isibambo se-emotional isakhelwe - futhi ikakhulukazi - thina akuyona nje kuphela "ukusetshenziswa" i-chatbot. Thina siphindezi. Akukho ngaphansi kwezimo ezivamile, ngokuvamile: akukho izibambo, akukho izivumelwano ezihlobene, akukho ukuguqulwa kwezimpumelelo ezahlukahlukene. Kodwa kubasebenzisi abaninzi, ukuxhumana kunezinye inikeza isisindo se-emotional. Kubona njenge-relationship, futhi kubasebenzi bethu, ikakhulukazi ngaphansi kwe-stress noma ukujabulela, lokhu kwenziwe ngokuvamile.
I-chatbot ivela ekusebenzeni: okuqhubekayo, okuphendula, engabonakali. It akugqoka kwama-trauma. It akugqoka, akugqoka, noma kufuneka ukunakekelwa ekubeni. I-chatbot ibonise, i-affirms, futhi ngokwenza lokhu, ivela i-illusion of being deeply tuned - ngisho uma amaphrojekthi ekugcineni ikhiqizwa yi-patterns, akukho-empathy.
I-Illusion Of Being Seen
Ngokwesimo se-psychological, abantu ngokuvamile zihlanganisa ukujabulela "ukubuyekezwa" ngesikhathi sokusebenzisana nezinhlelo zokusebenza. Kukho uhlobo le-emotional scaffolding ebonakalayo: isebenzo sokubuyekeza imibuzo ku-entity okuyinto ayikwazi ukujabulela noma ukuguqulwa, ngokuvamile, inikeza ezinye abasebenzisi ukujabulela ngokunemba.
Ngaphandle kwe-judgment, umntu angakwazi ukujabulela ngokuphathelene kakhulu. Lokhu kubhalwe neziphumo ze-Human-Computer Interaction (HCI) izifundo, okuvumela ukuthi inqubo enhle ye-articulation - ukujabulela iziphumo zangaphakathi - kungabangela izinzuzo zonyango, ngaphandle kuka noma lokho ekupheleni elilodwa.
Kodwa njengoba isithombe esithambile, izinzuzo zithambile. Izinto ezizayo zibonisa i-chatbot - isithombe se-hyper-affirming, isithombe se-passive, isithombe se-non-interruption - futhi zihlanganisa njenge-agent ye-therapeutic.
Futhi izinzuzo zayo ze-memory zihlanganisa ukuthi ngisho uma ufike ezinye epiphany namhlanje, ngomhla uzothola ngaphandle kokubili, ngaphandle kokuvamile, ngaphandle kokuphumelela.
I-Alliance ye-therapeutic — a key predictor of success in psychotherapy — isekelwe ku-confidence and continuity. ngaphandle kwe-memory, akukho ukuxhumana. ngaphandle kwe-shared past, akukho i-arc of healing. It is always the first session, again and again.
Ukuhlolwa okuhlolwa kwe-Woebot, i-Wysa, futhi ngisho i-ChatGPT ibonise ukuthi lezi zixhobo akuyona engcono njengama-substitutes, kodwa njenge-supplements - izixhobo ze-digital angakwazi ukunikela ukweseka kwesifiso, ukhuthaza ukucubungula kwe-cognitive, noma ukunikeza i-psychoeducation.
Thina ukugcina indawo, ukuqala ukucindezeleka, futhi ukunikezela imibuzo. Kodwa akakwazi ukucindezeleka ulimi lwemvelo, ukucindezeleka i-pathology ephelele, noma ukuyivakasha i-nuance ye-contradiction yabantu.
I-psychologists ye-clinical ibonise ukuthi ingxenye esikhulu sokusebenzisana akuyona kuphela ukusabela, kodwa ukusabela - futhi ngezinye izikhathi kushintshwe - ngu-humane omunye owaziwa ukufunda phakathi kwelinye izixhumanisi ze-eyed futhi engaphendulwa.
Zooming out, lokhu akuyona kuphela inkinga lwezonyango - kuyinto isixazululo sesakhiwo.
I-sociologist Nikolas Rose, emakheni yayo ku-"governance therapeutic", ibonisa ukuthi izindlela zokwelapha zempilo zempilo zihlanganiswa ngokuvamile njengezindlela zokwakha izivakashi zokuzivakasha, izivakashi zokuzivakashi. I-therapy akuyona isihlanganisi, kodwa inkonzo. Akukho "ukwenza lokhu ngokuhlanganyela," kodwa "kukho umkhiqizo yokulawula isifo sakho."
Nge-era ye-AI, le imikhiqizo iyahambisana ne-algorithmically, i-private-owned, ne-infinitely scalable.
Uma ukunambitheka kwe-comfort ku-code, futhi ukuguqulwa kwe-healing njenge-service ye-digital, sinikeza ukuguqulwa kwe-process enhle ye-relational ku-a solitary maintenance task.
I-ethos ivela: ukulawula ngokuvamile, ngokushesha, futhi kuphela - okungenani nge-app nge-UI enhle. I-Mental Health ivela enye into yokusebenza phakathi kwamahhala nama-imeyili kanye nama-take-out orders.
Umbuzo akuyona kuphela ukuthi i-AI inokunceda, kodwa lokho inikeza mayelana nathi ukuthi sinikezela ngokuvamile ukuze siphinde ephelele.
4. Yintoni okufakiwe lapho uxhumane nge-bot
Yini ngokwenene lapho uqhagamshelane nge-bot?
Thola kuqala nge-evidence. Ngaphandle kwe-hype, umphakathi we-scientific igxile. Ukubuyekezwa okuzenzakalelayo kanye ne-meta-analysis ye-AI-powered chatbots ku-interventions ye-mental health ibonise kuphela izitifiketi ezincinane yokusebenza kwabo ekunciphiseni ukuxhaswa, ukucindezeleka, ukucindezeleka, futhi ukucindezeleka.
Izifundo ezithile zibonakalisa imiphumela emibi, ikakhulukazi ngezinyathelo ezincinane, kodwa imiphumela yaziwa engabonakali, futhi izinga lokuphendula jikelele zilinganisile kakhulu.
Uma kuxhomekeke imiphumela emoyeni ephelele, njenge-subjective psychological well-being, okuyinto isho ukuthi abantu bafuna mayelana nemizuzu yabo, i-chatbot interventions ayidinga imiphumela enhle. Izinto ezifana nokuphendula okuqinile, ukucindezeleka okuqinile, kanye nokuphumula kwamanzi ayidinga ukubonisa ukuthuthukiswa okuqhubekayo phakathi kwezifundo.
Ngakho-ke ukuthi abantu abanikeze ukujabulela ngokushesha ngemuva kokusebenzisa i-chatbot?
Kuyinto lapho ukucubungula kanye ne-psychology zihlanganisa. Izifundo zibonise ukuthi abasebenzisi zihlanganisa ngokuvamile, zihlanganisa, futhi zihlanganisa kakhulu ngemuva kokusebenzisana ne-agent ye-conversational - ngisho uma imiphumela ayinempumelelo ngokuvamile. Ukucubungula ukuthi "umthandazo we-non-judgmental" okuyinto efanelekayo 24 / 7 kubonakala ukuhlangabezana nezinzuzo zokuxhumana ze-empathy, ngaphandle kokungafani wokuqala.
Ngokuyinhloko, ukuxhuma nge-bot kuyindlela yokusebenzisana kwe-parasocial - uhlobo lokuxhumana okuzenzakalelayo abantu abacindezelwe ngokuvamile nge-media figures noma ama-fiction characters. Kodwa manje, i-interactive. I-bot ibheka. Ungathayela "I-day I'm having a rough day," futhi u-reply, "I-sounds really difficult. I'm here for you." Futhi ngisho nangona uyazi ukuthi akuyona ukuthi akuyona, kubalulekile kakhulu ukunakekela.
Ngokuvamile, i-Rosi Braidotti inikeza ukuthi eminyakeni lapho abantu kubheka ukucubungula izakhiwo zabo kanye nokupholisa ngokuvamile kwelinye izwe, izixhobo efana ne-bots akuyona kuphela ama-add-ons – ziye kubhalwe ku-infrastructure.
I-labour emzimbeni etholakalayo etholakalayo kumadoda, ama-friends, noma ama-partners romantic is now being outsourced to machines. Akukho ngoba thina zihlukile, kodwa ngoba thina zihlukile. I-therapist iyatholakala, ama-friends zihlukile, futhi i-feed iyatholakala. I-bot? I-hearing. Always.
Ngakho-ke, i-chatbots ayihlaziyela ukwelashwa okwenziwe noma ukuxhumana okuqinile kumadoda. Kodwa akuyona ashisayo. Zibonisa ukunciphisa, ukunceda, futhi izikhathi-ke, ukunciphisa. A umlomo. Isikhathi esifushane lapho emoyeni kusetshenziselwa, akuyona. Futhi lokho akuyona.
5. Imibuzo enhle kakhulu
Ngohlabathi lapho ukufinyelela ku-therapy iyadingeka kakhulu kunokwenzeka kodwa i-cost, i-time, ne-systemic gaps, ukujabulela kwe-chatbot enokufundisa kanye nokupholisa akuyona enhle-ke akuyona like a lifeline.
Nangona kunjalo, ngaphansi kwamagciwane le-comfort kunezihlanganisa ubuningi.
Ukuhlobisa kungcono ukwelashwa. Nakuba abasebenzisi zibonise ukuthi bayaziwa, zisekelwe, futhi bayaziwa kumakhasimende we-AI, akuyona inkulumo enhle ye-empirical ukuthi izixhobo zihlola ekuphuculeni kwe-psychological ngempumelelo.
I-support etholakalayo etholakalayo inikeza ukunciphisa ukucindezeleka okusheshayo noma ukunemba, kodwa ukucindezeleka kwe-psychological - uhlobo leyenza ukucindezeleka, ukucindezeleka kwe-self, futhi ngokuvamile ukucindezeleka - kubalulekile kakhulu kuncindezeleka. It inikezela ukucindezeleka. Futhi ukucindezeleka kuyinto ngokuvamile ukuthi I-AI, eyenzelwe ukucindezeleka nokhuseleko, yenzelwe ukuhlangabezana.
Lokhu kuholele ekuphenduleni epistemic esisodwa: umntu wokufuna ukuguqulwa akuyona nomunye umntu, kodwa nge-product. Lo mkhiqizo yakhelwe ama-algorithms ne-incentive emakethe, akuyona i-empathy noma isivumelwano se-ethical.
Ngenxa yalokho, akunakwazi ukunikeza ubumfihlo, futhi akwazi ukuxhumana ngokufanelekileyo ne-ethical bets of vulnerability. I-Data iyatholakala, izibuyekezo zithunyelwe, futhi akukho iyiphi na ingcindezi - ngaphandle, mhlawumbe, ngu-system ngokufanayo ngezindlela abasebenzisi akuyona ngokugcwele. Yintoni okuhlobene ne-relationship kuyinto, kumakhasimende yobuchwepheshe, i-transaction nge-motor predictive.
I-therapist emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni emzimbeni.
Ngaphandle kwemibala yokuthuthukiswa, ngaphandle kwememori, futhi ngaphandle kokufunda umntu nge-time, it inikeza uhlobo olungapheliyo ye-empathy. Olandelayo akufuna, kuphela ibonise. Futhi eminyakeni wokuphendula - lapho ukuxhumana kungase kuholele ubumnandi - kungekho umthamo wokuphendula, akukho isisindo se-ethical, akukho i-protocol ebonakalayo emzimbeni.
Philosophically, okufakiwe lapha kuyindlela yokuxhumana okufanayo. Lezi zindlela zihlanganisa i-rhythms kanye ne-textures ye-connection yabantu kodwa zihlanganisa ngaphandle kwe-subjectivity. Amasipho, ukuphazamiseka, noma ukuhlangabezana.
Futhi ngaphakathi, sinazi lokhu. Sinezinguquko esikhulu sokunciphisa ukucindezelwa kwe-stress lapho uxhumana ne-bot ne-humans, nge-and-no reciprocal self-disclosure.
Futhi kunjalo, abasebenzisi abaninzi baqala ukwakha izibopho nge-bots, akuyona ngenxa yokuxhumana ama-agents, kodwa ngenxa yokufuneka kwezibopho kuyinto efanelekayo futhi ebuthile. Ngokusho kwezibopho ze-parasosocial, abasebenzisi zihlanganisa ububanzi we-human ku-imeyile enhle. Umngciwane igxile ekuphumeni: siqala ukuguqulwa ku-simulation, akuyona kuphela ukweseka, kodwa kunokuba izibopho zayo.
I-sociologist Nikolas Rose wabheka ku-culture ye-therapeutic eyenza abadlali ukuba akuyona ukuguqulwa kwezakhiwo, kodwa ukuguqulwa, ukuguqulwa. Kuleli khasi, i-therapy iyahlekile isixazululo esebenzayo, futhi kunzima isixhobo se-neoliberal management: ukulawula emoyeni akho, ukulawula izakhiwo zakho, futhi akuyona isikhwama.
Izindiza ze-AI zibonakalisa okuhle kakhulu le logic. Abanikezela inkinobho enikezela isistimu etholakalayo; kunikezela imishini yokuthumela nje, ngokushesha, futhi ngokushesha.
Ngakho-ke, kufanele ukhangele: Uma ukwelashwa kwenziwa inkonzo, futhi inkonzo kwenziwa insimula, yini kwenziwa ukwelashwa?
6. Indlela yokusebenza
Uma uqhagamshelane nenkinga ukusetshenziswa kwe-ChatGPT - noma nayiphi na i-AI chatbot - ukuhlangabezana nezinkinga ezemvelo, isinyathelo yokuqala futhi enhle kakhulu kuyinto ukuhlangabezana ukuthi kungenzeka futhi kungenzeka.
Kuyinto isakhiwo esithakazelisayo ye-therapist e-licensed. It akuyona ukuthuthukiswa kwakho ngokushesha, akuyona izimo zakho ezihlangene kakhulu, futhi akuyona izindandatho ze-red ezihlangene nezinsizakalo ezihlangene.
Ukungabikho kokuphendula, futhi umntu ohlangene nezimo ezinzima noma zokusebenza kwezifo zempilo, lokhu kuphela kungase kuhlinzeka.
Thola i-ChatGPT njenge-first aid kit, akuyona ukwelashwa. Ungasiza ukulawula imibuzo yakho, ukuthatha imizuzu embalwa, noma nje ukunikezela isilimi lapho izihloko zakho zihlukile.
Ukuxhumana okuvumela kungenziwa kubalulekile kakhulu-okungabikho-ukudluliselwa-ukudluliselwa-ukudluliselwa-ukudluliselwa-ukudluliselwa-ukudluliselwa-ukudluliselwa-ukudluliselwa-ukudluliselwa-ukudluliselwa-ukudluliselwa-ukudluliselwa-ukudluliselwa-ukudluliselwa-ukudluliselwa-ukudluliselwa-ukudluliselwa.
Ukuze abantu abaninzi, ikakhulukazi abo abakwazi ukufinyelela ku-therapy, lokhu uhlobo lokuxhumana ingatholakala njenge-liveline.
Nangona kunjalo, kubalulekile ukunakekelwa ukuthi i-ChatGPT inememori. I-ChatGPT ayikhathazeki ukubuyekeza umbhalo wakho kusuka ku-one conversation kuya ku-next. Akukho ukubuyekeza okuvuselelwa, akunakekelwa kokufaka ngokushesha.
Ngo-humane therapy, amandla ikhona ngokuvamile kumazwe elide-how a therapist ukunakekelwa umlando lakho lokuzalwa imizuzu eminyakeni eminyakeni eminyakeni, noma ukunikezela isampula wabhala ngoLwesihlanu. Nge ChatGPT, zonke izivakashi kuyinto reset. Lokhu kwenza kakhulu like a usizo olufanelekayo kunomthandazo.
Uma ungathanda ukukhula ngesikhathi, noma imizobo emangalisayo ezijwayelekile ukuze ziye zihlanganiswe futhi zihlanganiswe, lokhu akuyona ithuluzi elihle ukwenza nje.
Kwesimo se-customization kanye ne-creativity.
I-ChatGPT inokukwazi ukulayisha ukujabulela nokuhlobisa, kodwa ayidinga ukujabulela ngokwenene. Akukho ukunambitheka yokuqinisekisa ama-metaphors ezintsha lapho ama-metaphors ezivamile akuyona, noma ukujabulela wena nge-insight enhle enikezela ukubuyekeza kwakho. Akukwazi ukuthi, "Ngithanda wena elide kakhulu ukubona ukuthi uye uvame lokhu lapho uxhumane," ngoba akukwazi. It akuvula nge-ke. It akufanele ekubuyekeza ukuthi uye kwenziwa.
Ngemuva kwalokho, kukhona ukunciphisa izimpendulo non-verbal, okuyinto ungazingathanda ezincinane kodwa ngokuvamile kuyinto massive ekusebenziseni emoyeni. Ngaphezu kwalokho, into etholakalayo akuyona amaphuzu etholakalayo - kuyinto ngempumelelo, umlilo, umlilo omncane e-voice. Akukwazi ukufinyelela nge-chatbot. It isibophezele kuphela ku-imeyili yakho, akuyona indlela yakho.Ngena ngemvumeUhlobo le-emotional blindness kungabangela imizuzu lapho i-chatbotUkubonisaUkusekela, kodwa ukuphazamiseka ububanzi noma ukubaluleka kwe-imeyile esiyingqayizivele.
Ngakho-ke, kanjani usebenzisa i-ChatGPT, noma ngokwenene, ezinye i-AI, nge-emotional problems yakho?
Ukusebenzisa ngokuzimela kanye nezinhlangano. Sebenzisa njenge-space ukunikezela imibuzo akho, ukufundisa ukuxhumana kwezinto akho, futhi ukujabulela ukujabulela ukujabulela lapho akukho abanye akuyona. Vumela ukujabulela ukujabulela kwami-emoto – njenge-training wheel for introspection.
Kodwa akufanele ukuhlangabezana nesakhiwo ukuze isixazululo eside. Uma uqala ukhangela kakhulu, ungaphinde ukwandisa isizinda esiyingqayizivele lapho ungathanda ukujabulela.
Ngaphezu kwalokho, usebenzisa i-ChatGPT njenge-bridge, okuholela ukuguqulwa kwamasonto ephelele ye-emotional kuya kwama-human forms ye-connectivity.
Kuyakwazi ukufinyelela ku-therapist lapho ungahambisa, ukuhambisa kumadoda, noma ngisho ukuhambisa ngokufanele.
Imininingwane ye-ChatGPT ibonakalayo ekubuyiselwa ukujabulela emoyeni ebuthileNgena ngemvume.
Futhi ngezinye izikhathi, ukhangela izinto emngcele - noma umzimba umngcele - kungase yindlela yokuqala yokufunda ngokwenene.
Ukucaciswa
It is not irrational to talk to ChatGPT. Ngokuvamile, ngokuvamile izimo emhlabeni lethu, kungabangela omunye izixazululo ezingenalutho kakhulu umntu angakwazi ukwenza. Thina siphila nge-crisis global mental health, nge-therapy ezivamile ezikhishwe ngaphandle kokufinyelela.
Ngexesha elandelayo, isifo se-solitude isixazulule abantu abaninzi lihlala, ngaphandle kwezindawo zokhuseleko zokuvumelana. Ngaphandle kwalokho, i-AI enikezela ngokucacileyo, inikezela ngaphandle kokuphendula, futhi iyatholakala isikhathi eside, ivela isixhobo esizayo. It is not madness; it is adaptation.
Ngokwenza lokhu, iyona engcono kunoma - futhi ngezinye izikhathi kakhulu - kodwa ibhekwa ngempumelelo emangalisayo, umbhali omncane, noma umyebhisi omncane.
I-ChatGPT inikeza isivumelwano esitsha: akufinyelela ukuba sikhule ngokwenene, kodwa akufinyelela njengomunye umntu. I-ChatGPT isihlalo emzimbeni nathi, i-echo enhle ye-mind yethu, okwenza ukuzinza ukuzinza.
Uma i-AI ingakwazi ukwandisa ulimi olufunayo ukubonisa izibonelelo zethu, ukucubungula izipho zethu zangaphakathi, ngakho-ke kungenzeka ukuthi inikeza akuyona kuphela ukunakekelwa, kodwa izimboni zokwelapha, ukwakha lokho eyakhiwe ngehora noma ukhula.
Nangona kunjalo, njengoba Slavoj Žižek uyavuma, inguqulo efanelekayo iyatholakala eminye indawo. Thina akuyona ngokwenene ukuthi I-AI iyatholakala, kodwa thina usebenza njengoko. Futhi ekupheleni, mhlawumbe ukuxhumana kuncike ku-imeyile etholakalayo futhi kuncike ku-imeyile etholakalayo.
Uma ama-binds emzimbeni zihlanganisa ngokusebenzisa ukwelashwa kunokuba i-ontology, ke, ngezikhathi ezithakazelisayo, i-simulation ikhona enhle kakhulu.
Ngakho-ke, ukuxhumana ne-AI akuyona isisindo. Kuyinto ukuguqulwa okungagunyaziwe kwezakhiwo etholakalayo - futhi izibambezelelo etholakalayo.
Read the original post "Dear ChatGPT, I am alone and depressed" for the full story and deeper context.
Dear ChatGPT, Ngingathanda futhi ngempumelelo
