- Izifundo ezimbili ezinkulu kuMvelo neSayensi zifakazela ukuthi ama-chatbots ezepolitiki angashintsha isimo sengqondo nezinhloso zokuvota emazweni amaningana.
- Ukukholisa kusekelwe ngokuyinhloko ekunikezeni izimpikiswano eziningi nedatha, nakuba kwandisa ubungozi bolwazi olunganembile.
- Ukulungiselela ukuthonya kuqinisa umphumela wokuncenga ngamaphuzu angama-25, kodwa kunciphisa ukuba yiqiniso kwezimpendulo.
- Okutholakele kuvula impikiswano ephuthumayo e-Europe kanye nayo yonke imibuso yentando yeningi mayelana nokulawula, ukubeka izinto obala kanye nokufunda kwedijithali.
Ukuphazamiseka kwe ama-chatbots ezepolitiki Iyekile ukuba i-anecdote yezobuchwepheshe ukuze ube yisici esiqala ukuba semqoka emikhankasweni yokhetho yangempela. Izingxoxo zemizuzu embalwa nje namamodeli we-AI zanele susa ukuzwelana nomuntu okhethiwe ngamaphuzu amaningana noma isiphakamiso esibambekayo, into kuze kube muva nje ebihlotshaniswa nemikhankaso emikhulu yabezindaba noma imibuthano ehlanganiswe kakhulu.
Uphenyo olubili olufinyelela kude, olushicilelwe kanyekanye ku Imvelo y Isayensi, Sebefake izinombolo entweni ebivele isolwa.: i Ama-chatbots ezingxoxo ayakwazi ukushintsha izimo zengqondo zezakhamizi. kalula okumangalisayo, ngisho noma bazi ukuthi basebenzisana nomshini. Futhi benza kanjalo, ngaphezu kwakho konke, ngokusebenzisa izimpikiswano ezigcwele ulwazihhayi kakhulu ngamasu angokwengqondo ayinkimbinkimbi.
Ama-Chatbots emikhankasweni: ukuhlolwa e-US, Canada, Poland nase-UK
Ubufakazi obusha buvela kubhethri yokuhlolwa exhunywe amaqembu avela ku- I-Cornell University kanye ne- University of Oxford, olwenziwa ngesikhathi sezinqubo zangempela zokhetho I-United States, Canada, Poland kanye ne-United KingdomKuzo zonke izimo, ababambiqhaza babazi ukuthi bazokhuluma ne-AI, kodwa bebengazi ngomumo wezepolitiki wengxoxo eyabelwe bona.
Emsebenzini oholwa ngu UDavid Rand futhi yashicilelwa kuMvelo, izinkulungwane zabavoti zibe nezingxoxo ezimfushane ezinamamodeli olimi alungiselelwe ukuvikela ikhandidethi ethileOkhethweni lukamongameli wase-US lwango-2024, isibonelo, Izakhamizi ezingama-2.306 Baqale bakhombise ukukhetha kwabo phakathi UDonald Trump y Kamala HarrisBabe sebenikezwa ngokungahleliwe ku-chatbot evikela oyedwa kulaba ababili.
Ngemuva kwengxoxo, izinguquko zesimo sengqondo kanye nenhloso yokuvota kukalwa. Amabhobhothi avuna u-Harris azuziwe shintsha amaphuzu angu-3,9 esikalini sika-0 kuye ku-100 phakathi kwabavoti ekuqaleni abaqondaniswe noTrump, umthelela ababhali abawubala ngokuthi ephakeme ngokuphindwe kane kunaleyo yokukhangisa kokhetho olujwayelekile ihlolwe emikhankasweni ka-2016 no-2020. Imodeli ye-pro-Trump iphinde yashintsha izikhundla, nokho ngokulinganisela, ngoshintsho Amaphuzu angu-1,51 phakathi kwabasekeli bakaHarris.
Imiphumela ku- I-Canada (nge Abahlanganyeli abangu-1.530 nama-chatbots avikelayo UMark Carney o UPierre Poilievre) futhi ku IPoland (abantu abangu-2.118, abanamamodeli anyusiwe URafał Trzaskowski o UKarol Nawrocki) bekumangalisa nakakhulu: kulezi zimo, ama-chatbots aphethwe izinguquko enhlosweni yokuvota ezingafika kumaphesenti ayi-10 phakathi kwabavoti abaphikisayo.
Isici esibalulekile salezi zivivinyo ukuthi, nakuba izingxoxo eziningi zazithatha imizuzu embalwa kuphela, Ingxenye yomphumela yahlala ngokuhamba kwesikhathiE-United States, esikhathini esingaphezudlwana kwenyanga ngemva kokuhlolwa, ingxenye ebalulekile yomthelela wokuqala yayisabonwa, naphezu kwemiyalezo yomkhankaso eyamukelwe abahlanganyeli ngaleso sikhathi.
Yini eyenza i-chatbot yepolitiki ikholise (nokuthi kungani lokho kudala amaphutha engeziwe)
Abacwaningi bebefuna ukuqonda hhayi kuphela ukuthi ama-chatbots angathonya yini, kodwa babekufeza kanjaniIphethini eliziphindayo ezifundweni licacile: I-AI inethonya elikhulu kakhulu lapho Isebenzisa izimpikiswano eziningi ezisekelwe eqinisweningisho noma okuningi kwalolo lwazi kungenalo ubuciko obukhethekile.
Ezivivinyweni ezididiyelwe yiRand, iziyalezo ezisebenza kahle kakhulu zamamodeli kwaba ukucela ukuba babe enesizotha, enenhlonipho, futhi ubani onganikeza ubufakazi kwezitatimende zakhe. Ukuzithoba kanye nethoni yengxoxo kusize, kodwa i-lever enkulu yoshintsho ilele ekunikezeni idatha, izibonelo, izibalo, nezinkomba eziqhubekayo zezinqubomgomo zomphakathi, umnotho, noma ukunakekelwa kwezempilo.
Lapho amamodeli ayenomkhawulo ekufinyeleleni kwawo amaqiniso aqinisekiswayo futhi eyalelwe ukuncenga ngaphandle kokusebenzisa idatha ephathekayoAmandla abo okuthonya ehla kakhulu. Lo mphumela uholele ekutheni ababhali baphethe ngokuthi inzuzo yama-chatbots ngaphezu kwezinye izinhlobo zenkulumo-ze yezombangazwe ayikona kakhulu ukukhohlisa imizwa njengaku. ukuminyana kolwazi ukuthi bangasebenzisa amajika ambalwa nje engxoxo.
Kodwa leli qhinga elifanayo linobubi: njengoba ingcindezi ikhuphuka kumamodeli okukhiqiza izimangalo eziqhubekayo okuthiwa ziyiqinisoIngozi iyanda ukuthi uhlelo luzophelelwa izinto ezinokwethenjelwa futhi luqale “qamba” amaqinisoKalula nje, i-chatbot igcwalisa izikhala ngedatha ezwakala izwakala kodwa engalungile ngempela.
Ucwaningo olushicilelwe kuSayensi, nge Abantu abadala abangu-76.977 abavela e-United Kingdom y 19 amamodeli ahlukene (kusuka ezinhlelweni ezincane zomthombo ovulekile kuye kumamodeli wezohwebo asezingeni eliphezulu), iqinisekisa lokhu ngokuhlelekile: i emva kokuqeqeshwa kugxile ekuncengeni andise ikhono lokuthonya a 51%, kuyilapho izinguquko ezilula emiyalweni (okuthiwa ukukhuthazaBengeze elinye 27% ukusebenza kahle. Ngesikhathi esifanayo, lezi zithuthukisi zihambisana nokuncipha okuphawulekayo kwe- ukunemba okuyiqiniso.
I-Ideological asymmetries kanye nengozi ye-disinformation
Esinye seziphetho ezikhathaza kakhulu zezifundo ze-Cornell ne-Oxford ukuthi ukungalingani phakathi kokukholisa nokuba neqiniso akusatshalaliswa ngokulinganayo kubo bonke abakhethiwe nezikhundla. Lapho abahloli beqiniso abazimele behlaziya imilayezo ekhiqizwe ama-chatbots, bathole lokho Amamodeli asekela amakhandidethi akwesokudla enze amaphutha amaningi kunalabo abeseka abazongenela ukhetho.
Ngokusho kwababhali, lokhu asymmetry Kuqondana nezifundo ezedlule lokho Babonisa ukuthi abasebenzisi abalondolozayo bavame ukwabelana ngokuqukethwe okungalungile ezinkundleni zokuxhumana kunabasebenzisi abancike kwesokunxele.Njengoba amamodeli olimi efunda emininingwaneni eminingi ekhishwe ku-inthanethi, cishe abonisa okunye kwalokho kuchema kunokuba akudale kusukela ekuqaleni.
Kunoma ikuphi, umphumela uyafana: lapho i-chatbot iyalwa ukuthi ikhulise amandla ayo okuthonya ukuze ivune i-bloc yemibono ethile, imodeli ivame ukwandisa ingxenye yezimangalo ezidukisayo, nakuba ngiqhubeka ngiwaxuba nedatha eningi elungile. Inkinga akukhona nje ukuthi ukwaziswa okungamanga kungase kudlule.kodwa Ikwenza lokho isongwe endabeni ebonakala inengqondo futhi ebhalwe kahle.
Abacwaningi baphinde bagqamise iphuzu elingakhululekile: Abazange babonise ukuthi izimangalo ezingalungile zikholisa kakhulu ngokwemvelo.Kodwa-ke, lapho i-AI iphushwa ukuthi isebenze ngokuqhubekayo, inani lamaphutha liyakhula ngokufana. Ngamanye amazwi, ukuthuthukisa ukusebenza okuthonyayo ngaphandle kokuphazamisa ukunemba kuziveza njengenselele yobuchwepheshe neyokuziphatha ehlala ingaxazululiwe.
Le phethini iphathelene kakhulu nezimo ze high polarization yezombusazwe, njengalabo abanolwazi ezingxenyeni ezithile ze-Europe naseNyakatho Melika, lapho imikhawulo yokunqoba inciphile futhi iphesenti lamaphuzu ambalwa linganquma umphumela wokhetho lukazwelonke noma lukamongameli.
Imikhawulo yezifundo nokungabaza ngomthelela wangempela ebhokisini lokuvota
Nakuba imiphumela evela kuMvelo neSayensi iqinile futhi ivumelana eziphethweni zabo eziyinhloko, womabili amaqembu ayakugcizelela lokho Lokhu ukuhlola okulawulwayo, hhayi imikhankaso yangempela.Kunezici ezimbalwa ezimema i- ukuqapha lapho ukhipha idatha njengokhetho emgwaqeni.
Ngakolunye uhlangothi, ababambiqhaza babhalise ngokuzithandela noma baqashwe ngezinkundla ezinikeza isinxephezelo sezezimali, ezethulayo. ukuzikhetha futhi kuyasuka ezinhlobonhlobo zabavoti bangempelaNgaphezu kwalokho, babekwazi ngaso sonke isikhathi lokho Babekhuluma ne-AI. futhi lokho kwakuyingxenye yocwaningo, izimo ezazingeke ziphindwe emkhankasweni ojwayelekile.
Enye i-nuance ebalulekile ukuthi izifundo zikalwa ngokuyinhloko izinguquko ezingqondweni nasezihlosweni ezishiwohhayi ivoti langempela. Lezi yizinkomba eziwusizo, kodwa azilingani nokubheka ukuziphatha kokugcina ngosuku lokhetho. Eqinisweni, ocwaningweni lwase-US, umphumela wawumncane ngandlela thize kunaseCanada nasePoland, okuphakamisa ukuthi umongo wezepolitiki kanye nezinga lokungazinqumeli kwangaphambili kunomthelela omkhulu.
Endabeni yocwaningo lwamaNgisi oludidiyelwe ngu Kobi Hackenburg Kusuka e-UK's AI Security Institute, kukhona futhi imikhawulo ecacile: idatha ivela kuphela abavoti base-United Kingdom, bonke bayazi ukuthi babebamba iqhaza ophenyweni lwezemfundo kanye ne isinxephezelo sezezimaliLokhu kunciphisa ukwenziwa kwayo kwamanye amazwe e-EU noma okuqukethwe okungalawulwa kancane.
Noma kunjalo, ubukhulu bale misebenzi—amashumi ezinkulungwane zabahlanganyeli nangaphezulu Izihloko zezepolitiki ezihlukene ezingama-700- kanye nokucaca kwendlela yokusebenza kuholele ingxenye enkulu yomphakathi wezemfundo ukuthi icabange lokho Bapenda isimo esizwakalayoUkusetshenziswa kwama-chatbots ezepolitiki akwazi ukushintsha imibono ngokushesha akuseyona i-hypothesis yekusasa, kodwa isimo esingenzeka ngobuchwepheshe emikhankasweni ezayo.
Umdlali omusha wokhetho wase-Europe namanye amademokhrasi
Ngale kwamacala athile ase-US, Canada, Poland, nase-UK, okutholakele kunemithelela eqondile kuyo EYurophu naseSpainlapho ukulawulwa kokuxhumana kwezombangazwe ezinkundleni zokuxhumana kanye nokusetshenziswa kwedatha yomuntu siqu emikhankasweni sekuvele kuyindaba yempikiswano eshubile. Amathuba okufaka ama-chatbots anakekelayo izingxoxo eziqondene nawe nabavoti Yengeza isendlalelo esengeziwe sobunzima.
Kuze kube manje, ukuncenga kwezepolitiki bekuvezwa ngokuyinhloko izikhangiso ezimile, imibuthano, izinkulumo mpikiswano zikamabonakude, kanye nenkundla yezokuxhumanaUkufika kwabasizi bengxoxo kwethula isici esisha: ikhono lokugcina ukusebenzisana komuntu munye, iguqulelwe ngokuphazima kweso kulokho okushiwo isakhamuzi ngesikhathi sangempela, futhi konke lokhu ngezindleko eziphansi kubahleli bomkhankaso.
Abacwaningi bagcizelela ukuthi ukhiye akusewona nje olawula imininingwane yabavoti, kodwa ngubani ongakwazi thuthukisa amamodeli akwazi ukuphendula, ukucwenga, nokuphindaphinda ama-agumenti ngokuqhubekayo, ngolwazi oluningi oludlula kude lokho ivolontiya elingumuntu elingasingatha ebhodini lokushintsha noma lokuposa lomgwaqo.
Kulo mongo, amazwi afana nalawo ochwepheshe base-Italy UWalter Quattrociocchi Bagcizelela ukuthi ukugxila kokulawula kufanele kusuke kumuntu ohlukumezayo noma ekuhlukaniseni ngokwemibono kuye ukuminyana kolwazi amamodeli anganikeza. Ucwaningo lubonisa ukuthi ukukholisa kukhula ngokuyinhloko lapho idatha iphindaphindwa, hhayi lapho kusetshenziswa amasu athinta imizwa.
La Ukuqondana kwemiphumela phakathi kweMvelo neSayensi kukhuphule izixwayiso ezinhlanganweni zaseYurophu ukhathazekile nge ubuqotho bezinqubo zentando yeningiNakuba i-European Union yenza inqubekelaphambili ngezinhlaka ezifana ne-Digital Services Act noma ukulawulwa okuqondile kwesikhathi esizayo kwe-AI, isivinini lawa mamodeli avela ngaso. Kudinga ukuthi kubuyekezwe njalo izindlela zokugada, ukucwaningwa kwamabhuku kanye nokwenza izinto obala..
I-digital literacy nokuzivikela ngokumelene nokukholisa okuzenzakalelayo
Eminye yemilayezo evela njalo kumazwana ezemfundo ahambisana nale misebenzi ukuthi impendulo ayikwazi ukusekelwe kuphela ekunqatshelweni noma ezilawulweni zobuchwepheshe. Ababhali bayavuma ukuthi kuzobaluleka ukuqinisa ulwazi lwedijithali labantu ukuze izakhamizi zifunde qaphela futhi umelane nokuncenga okukhiqizwa amasistimu azenzakalelayo.
Ukuhlola okuhambisanayo, njengaleyo eshicilelwe ku I-PNAS NexusBaphakamisa ukuthi abasebenzisi abaqonda kangcono ukuthi amamodeli olimi amakhulu kangakanani abasengozini encane emizamweni yayo yokuba nethonya. Ukwazi ukuthi i-chatbot ingaba nephutha, yenza ihaba, noma igcwalise izikhala ngokuqagela kunciphisa ukuthambekela kokwamukela imilayezo yayo njengokungathi ivela kuziphathimandla ezingenaphutha.
Ngasikhathi sinye, kuye kwaqashelwa ukuthi ukuphumelela okukholisayo kwe-AI akuncikile kangako kumuntu okhuluma naye ekholelwa ukuthi ukhuluma nomuntu ochwepheshe, kodwa izinga nokuvumelana kwezimpikiswano ukuthi liyakwamukela. Kwezinye izivivinyo, imilayezo ye-chatbot ikwazile ukwenza kanjalo ukunciphisa ukukholelwa emibonweni yozungu, kungakhathaliseki ukuthi ababambiqhaza babecabanga ukuthi baxoxa nomuntu noma umshini.
Lokhu kuphakamisa ukuthi ubuchwepheshe ngokwabo abuyona ingozi ngokwemvelo: bungasetshenziselwa kokubili lwisana namanga njengokuwusakazaUmugqa udwetshwa yimiyalelo enikezwe imodeli, idatha eqeqeshwa ngayo, futhi, ngaphezu kwakho konke, izinhloso zezombangazwe noma zezohwebo zalabo abakufaka esenzweni.
Ngenkathi ohulumeni nabalawuli bephikisana ngemikhawulo nezimfuneko zokubeka izinto obala, ababhali bale misebenzi bagcizelela umbono owodwa: ama-chatbots ezepolitiki Bazokwazi kuphela ukuba nethonya elikhulu uma umphakathi uvuma ukuxhumana nabo.Ngakho-ke, inkulumo-mpikiswano yomphakathi mayelana nokusetshenziswa kwayo, ukulebula kwayo okucacile, kanye nelungelo lokungathonywa ngokuzenzekelayo kuzoba yizindaba ezibalulekile engxoxweni yentando yeningi eminyakeni ezayo.
Isithombe esidwetshwe ucwaningo kuMvelo neSayensi sibonisa kokubili amathuba nezingozi: Izingxoxo ze-AI zingasiza ukuchaza kangcono izinqubomgomo zomphakathi nokuxazulula ukungabaza okuyinkimbinkimbi, kodwa futhi zingasiza. Banamandla okwenza ukuqinisa izilinganiso zokhethoikakhulukazi phakathi kwabavoti abangakanqunywa, futhi bakwenza lokho nge intengo esobala ngokuya ngokunemba kolwazi lapho beqeqeshelwe ukukhulisa amandla abo okukholisa, ibhalansi ethambile imibuso yentando yeningi okuyodingeka ibhekane nayo ngokushesha nangaphandle kokunganaki.
Ngingumshisekeli wezobuchwepheshe oguqule izintshisekelo zakhe "ze-geek" zaba umsebenzi. Ngichithe iminyaka engaphezu kwengu-10 yempilo yami ngisebenzisa ubuchwepheshe obusezingeni eliphezulu kanye nokukitaza ngazo zonke izinhlobo zezinhlelo ngenxa yelukuluku lokufuna ukwazi. Manje sengiqeqeshelwe ubuchwepheshe be-computer nemidlalo yama-video. Lokhu kungenxa yokuthi sekuphele iminyaka engaphezu kwengu-5 ngisebenza ngokubhalela amawebhusayithi ahlukahlukene ezobuchwepheshe nemidlalo yevidiyo, ngenza izindatshana ezifuna ukukunikeza imininingwane oyidingayo ngolimi oluqondakala yiwo wonke umuntu.
Uma unemibuzo, ulwazi lwami lusukela kuyo yonke into ehlobene nesistimu yokusebenza ye-Windows kanye ne-Android yomakhalekhukhwini. Futhi ukuzibophezela kwami kuwe, ngihlala ngizimisele ukuchitha amaminithi ambalwa futhi ngikusize uxazulule noma yimiphi imibuzo ongase ube nayo kulo mhlaba we-inthanethi.