Agentic AI ya rusa zaton da ke bayan kowane hasashen fuskantar ayyuka

Hasashen kasuwar kwadago sun dauka cewa sarrafa kansa (automation) yana cire ayyuka daya bayan daya kuma yana barin sana'ar a tsaye. Agentic AI, wanda ke kammala daukacin tsarin aiki (workflows), yana soke wannan zaton, kuma an gina wani sabon ma'auni na hadin gwiwa, Agentic Task Exposure (ATE), don auna abin da tsofaffin tsare-tsare ba za su iya ba.

Takaitaccen Bayani na Zartarwa

Hasashen farko na sarrafa kansa, ciki har da wadanda aka gina a kan tsarin ayyuka na Autor-Levy-Murnane, suna da zato guda daya: sana'ar tana tsira a matsayin kwaryar hadin gwiwa a kewayen sauran ayyukan dan Adam.

Tsarin agentic AI yana cika sharudda hudu: suna karbar bayanin manufa, suna zabar kayan aiki da kansu kuma su kira su, suna kiyaye yanayi (state) a cikin matakan tunani, kuma suna samar da cikakken sakamakon aiki ba tare da sa hannun mutum a matakan tsakiya ba.

Makin ATE yana hada makin iyawar AI, abubuwan da ke rufe tsarin aiki da saurin karbuwa na logistic. Masu sharhi kan bashi (credit analysts), alkalai da kwararrun dorewa (sustainability specialists) sun kai makin ATE na 0.43-0.47.

Takardar ta yi amfani da ATE a kan lambobin SOC 236 a cikin manyan rukunin sana'o'i shida a yankunan fasaha biyar na Amurka, kuma ta bayyana kanta a matsayin aikin farko na gwaji mai yankuna da yawa na tsarin fuskantar ayyuka wanda ke tunkarar tsarin agentic AI musamman.

Tare da maye gurbin ayyuka, takardar ta gano rukunin sana'o'i goma sha bakwai masu tasowa, wadanda suka tattara a kan hadin gwiwar mutum da AI, shugabancin AI da ayyukan AI na musamman a fannoni daban-daban.

Hasashen da suka wuce gona da iri #

Robobin masana'antu sun maye gurbin ma'aikatan kere-kere. Amma ba su yi hakan a kan lokacin ko kuma a ma'aunin da hasashen farko suka nuna ba. Wannan jimlar ya kamata ta sa duk wanda ke karanta hasashen kasuwar kwadago a matsayin sana'a ya dan yi taka-tsantsan, domin tana bayyana wani tsari: alkiblar ta yi daidai amma girman abin ya kuskure.

Lambar da ta fi shahara a wannan fannin ta fito ne daga Frey da Osborne, wadanda suka nemi kwararru su ba wa sana'o'in Amurka 702 maki a kan "damar sarrafawa da kwamfuta" gaba daya. Sakamakonsu shi ne kashi 47% na sana'o'i suna cikin babban hadari. Lamba ce mai ban mamaki, kuma ta yadu sosai.

Daga nan sai Arntz, Gregory da Zierahn suka nuna abin da ke faruwa idan binciken ya sauko zuwa matakin daidaikun ayyuka (task level) kuma ya yi la'akari da bambance-bambancen ayyuka a cikin sana'o'in da ake ganin suna fuskantar barazana. Kiyasin babban hadari ya sauko zuwa kashi 9-14% a kasashen OECD. Tazarar da ke tsakanin 47% da 9-14% ba sabani ba ne na kuskuren lissafi. Yana nuna zabi ne game da na'urar bincike (unit of analysis), kuma wannan zabin yana motsa amsar sau uku ko fiye.

Tushen bayanin ya fito fili ya nuna cewa muhawarar da ke tsakanin kiyasin matakin sana'a da matakin ayyuka ba a gama warware ta ba. Abin da fannin ya cimma shi ne matsaya mai amfani: rarrabawa a matakin ayyuka yana da muhimmanci, kuma lakabin matakin sana'a ya yi dunkulewar da ta wuce kima. Wannan matsayar ita ce harsashin da sauran hujjar ke tsaye a kai. Haka kuma ita ce harsashin da agentic AI ke dora wa nauyi mai tsanani.

Yadda tsare-tsaren ayyuka ke aiki #

Don ganin inda nauyin ke fitowa, yana da amfani a fayyace abin da al'adar binciken ayyuka ke ikirari.

Autor, Levy da Murnane sun raba ayyukan sana'a a kan ginshikai biyu: na yau da kullun (routine) da wanda ba na yau da kullun ba (non-routine), da kuma na tunani (cognitive) da na hannu (manual). Hujjarsu ita ce sarrafa kansa yana maye gurbin na yau da kullun yayin da yake cike gurbin wanda ba na yau da kullun ba. Tunani ne mai tsafta, kuma ya ba wa ayyukan bincike na baya kalmomin da za su yi amfani da su wajen magana kan wadanne bangarori na aiki ne injin zai iya dauka.

Acemoglu da Restrepo sun gina tsari na doka a kan wannan. Suna kwatanta abin da ake samarwa (output) a matsayin sakamakon jerin ayyuka masu ci gaba da aka raba tsakanin jari da kwadago. A cikin tsarinsu, kason kwadago yana tashi ko yana sauka gwargwadon ko kirkirar sabbin ayyuka, wanda suke kira hanyar dawowa (reinstatement channel), ta fi saurin maye gurbin tsofaffin ayyuka. A bisa bin nasu tsarin, ana rubuta kason kwadago a cikin karin daraja a matsayin sL = W * L / Y = f(|NL| / |N|), inda W yake matsayin albashi, L shi ne daukar aiki kuma Y shi ne abin da aka samar. Karfi guda biyu suna jan juna, kuma duk wanda ya yi nasara shi ke kayyade wanda zai dauki ribar.

Wasu kokarin auna abubuwa guda biyu suna da muhimmanci a nan domin sun juya tsarin zuwa lambobi na matakin sana'a. Felten, Raj da Seamans sun gina makin AI Occupational Exposure (AIOE) ta hanyar hada ci gaban iyawar AI da aka wallafa zuwa ga rarrabuwar kwarewa ta O*NET. Eloundou da abokan aikinsa sun gina wani ma'auni makamancin haka wanda ya mayar da hankali kan GPT-4,ដោយ hada bayanan ma'aikatan taron jama'a (crowd-worker annotations) tare da gwajin kai-da-kai na GPT-4 a matakin ayyuka.

Babban abin da suka gano shi ne kashi 80% na ma'aikatan Amurka suna rike da ayyukan da akalla kashi 10% na ayyukansu suna fuskantar tasirin LLM. Fuskantar mafi nauyi ta tattara ne a tsakanin sana'o'i masu babban albashi da ke bukatar digiri. Wannan sakamakon ya ci karo da tsohon tunanin cewa sarrafa kansa yana fara dukan ayyukan yau da kullun masu karamin albashi ne da farko, kuma yana bayyana dalilin da ya sa tambayar ma'aikatan ofis (white-collar) ta zama mai gaggawa.

Matsalar rashin gani da kowane tsari ke tarayya a ciki #

Tushen bayanin yana yin ikirari mai kaifi game da duk wannan aikin. Kowane tsari da aka ambata yana da matsalar rashin gani ta tsarin gini: suna kwatanta sarrafa kansa a matsayin tsince daidaikun ayyuka yayin da sana'ar ke zama daram. Kusan dukkan hasashe, ciki har da wadanda aka gina a kan tsarin ayyuka na Autor-Levy-Murnane, suna tarayya a cikin zaton cewa sarrafa kansa yana daukar ayyuka daya bayan daya kuma sana'ar tana tsira a matsayin kwaryar hadin gwiwa a kewayen sauran aikin mutum.

Wannan zato ne game da tsari, ba game da fasaha ba. Ba ya cewa komai game da yadda injin din yake da kwarewa. Yana cewa komai yawan ayyukan da injin din ya dauka, har yanzu dole ne wani ya hada sauran wuri guda.

Mataimakin lauya (paralegal) yana sa batun ya fito a zahiri. Binciken fuskantar ayyuka na iya nuna kashi 80% na ayyukan mataimakin lauya a matsayin wadanda ke fuskantar LLM amma duk da haka ya yanke shawarar cewa sana'ar tana cikin aminci, domin hada wadannan ayyukan wuri guda yana bukatar hukuncin dan Adam. Kammalawar ta biyo bayan zaton kwarya. Ayyukan suna fuskantar barazana, amma dinkin da ke tsakaninsu ba ya fuskanta, don haka aikin yana nan daram.

Sannan babban matakin tushen bayanin: cikas din hadin gwiwa wanda ya kare matsayin mataimakin lauya yanzu yana cikin sashen tsare-tsare (planning module) na agent din. Idan hakan ya tabbata, tunanin da ya kiyaye mataimakin lauya cikin aminci yana daina aiki. Kariyar ta fito ne daga hadin gwiwar. Ita kuma hadin gwiwar ita ce ainihin abin da tsarin agentic yake yi.

Ina ganin wannan shi ne tunani mafi amfani a cikin takardar, kuma ya cancanci a dauke shi a kan nasa sharuddan. Tsofaffin tsare-tsaren ba su yi kuskure ba game da ayyukan. Sun yi kuskure ne game da abin da ke ragewa bayan ayyukan sun tafi. Ma'aunin da ke ba wa ayyuka maki daya bayan daya ba zai iya ganin tsarin da ke tafiyar da jerin ayyukan a matsayin abu guda ba.

Abin da ake lissafawa a matsayin agentic #

Idan hujjar ta dogara ne akan bambanci, bambancin yana bukatar ma'ana. Tushen bayanin yana ba da ma'ana mai sharudda hudu. Tsarin agentic AI shi ne wanda:

yake karbar bayanin manufa (goal specification) maimakon bayanin aiki (task specification);
yake zabar kayan aiki da kansa kuma ya kira su don cimma manufar;
yake kiyaye yanayi (state) a cikin matakan tunani da yawa; kuma
yake samar da cikakken sakamakon aiki ba tare da sa hannun mutum a matakan tsakiya ba.

Wannan jeri ne da ya kamata a rike shi a matsayin jeri, domin sharuddan suna rabe daya bayan daya kuma ana iya gwada kowannensu. Tsarin da ke amsa tambaya (prompt) yana faduwa a na farko. Tsarin da ke bukatar mutum ya amince da kowane matakin tsakiya yana faduwa a na hudu.

A kalmomi masu sauki, agent yana karbar manufa, yana zabar kayan aiki, yana aiwatar da shiri mai matakai da yawa, yana duba nasa sakamakon kuma yana isar da kammalallen sakamakon aiki. Tsarin agentic yana jera kiran kayan aiki, yana kiyaye tunani mai dorewa a tsakanin matakai, yana tsara jerin ayyuka da yawa kuma yana gyara kansa lokacin da sakamakon tsakiya ya gaza cin jarabawar inganci.

Bayanan gyara kai-da-kai yana da muhimmanci fiye da yadda yake nunawa a farko. Bututun aiki (pipeline) da ke karyewa idan mataki daya ya gaza har yanzu yana bukatar mutum a kusa. Tsarin da ke lura da gwajin inganci da ya gaza kuma ya sake gwadawa ya riga ya shanye wani bangare na sa idon da a baya yake ba da hujjar kasancewar mutum a wurin.

Takardar ta bayyana abin da take nazari ta wannan hanyar: agents na AI masu cin gashin kansu masu iyawar kammala daukacin tsarin aikin sana'a maimakon daidaikun ayyuka. Kalmar "daukacin" ita ke dauke da nauyin. Dukkan hujjar tana kan tazarar da ke tsakanin aiki daya da kuma cikakken tsarin aiki.

Me ya sa lokacin yake yanzu #

Tushen bayanin yana ba da nau'ikan shaidu guda biyu cewa tambayar ba ta zato ba ce. Ta farko ta shafi iyawa. Nau'ikan samfura irin na GPT-4 sun kai ko sun wuce mizanin makin cin jarabawa a kan jarabawar lauyoyi (bar exam) da Jarabawar Lasisin Likitoci ta Amurka, kuma sun zo cikin 'yan maki kadan na kashi dari don cin jarabawar CFA Level I. Wadannan jarabawar ba da shaidar kwarewa ne ga sana'o'in da ke takaita shiga musamman saboda aikin yana da wahala.

Ta biyun ta shafi girkawa a fagen aiki (deployment). Manyan dillalan fasahar kamfanoni, ciki har da Amazon, Microsoft, Salesforce, ServiceNow da Workday, sun bayyana girka tsarin agentic a fagen samarwa (production deployments) da ke taba tsarin aikin ma'aikatan ofis a cikin rahotannin kudaden shiga na 2024 da gabatarwa ga masu zuba jari. Wadannan girkawa ne na fagen samarwa na gaske, ba gwaje-gwajen bincike ba, kuma an sanar da su ne ga masu zuba jari maimakon malaman jami'a.

Tushen bayanin yana karanta wadannan sanarwar a matsayin irin wadanda kamfanoni ke yi kawai lokacin da girkawar ta riga ta fara shafar shirin adadin ma'aikata (headcount planning). Tushen bayanin yana gabatar da hakan a matsayin fassara, ba ma'auni ba, kuma ya kamata a dauke shi a haka. Babu wani abu a cikin rikodin da aka ba ni da ke ba da rahoton alkaluman adadin ma'aikata kai tsaye. Amma kammalawar tana da ma'ana: kamfanoni ba sa cika gabatar da wani tsarin da suka girka a gaban masu zuba jari sai dai idan yana da tasiri ga tushen kashe kudi.

A hada biyun wuri guda kuma hujjar neman sabon ma'auni tana fitowa fili. Iyawa kadai ba za ta tilasta batun ba, haka ma 'yan gwaje-gwaje kadan. Amma iyawar da tuni ta ketare matakan kwararru, a cikin tsare-tsaren da tuni suke aiki a fagen samarwa, tana tilastawa.

Makin ATE: abin da yake aunawa da abin da ba ya aunawa #

Martanin takardar shi ne makin Agentic Task Exposure. Ma'auni ne na hadin gwiwa wanda ke fadada tsarin maye gurbin ayyuka na Acemoglu-Restrepo don yin la'akari da rufe tsarin aiki daga farko zuwa karshe, ma'aunin iyawar agentic AI na yanzu da kuma saurin karbuwa a kamfanoni. Makin yana auna matakin da wata sana'a o take fuskantar maye gurbi daga farko zuwa karshe ta hanyar tsarin agentic AI na yanzu.

Abubuwa uku ne ke shiga ciki: makin iyawar AI, abubuwan da ke rufe tsarin aiki (workflow coverage factors) da saurin karbuwa na logistic. Abin da ke rufe tsarin aiki shi ne bangaren da ke tunkarar matsalar rashin ganin kai tsaye. Tushen bayanin ya ce abin da ya bata a ayyukan baya shi ne ma'auni na ko daukacin tsarin aikin wata sana'a yana zaune a cikin iyakar aikin (operational envelope) na agentic AI, kuma ya tsara hakan a hukumance a matsayin workflow coverage factor. Tsofaffin makin suna tambaya ko wani aiki yana fuskantar barazana. Wannan yana tambaya ko daukacin tsarin aikin yana shiga cikin abin da agent zai iya yi.

An bayyana karin bangarori biyu na ginin. Kalmar wo,t ita ce muhimmancin aiki t a cikin sana'a o da aka auna da nauyi, wanda aka ciro daga makin muhimmancin ayyuka na O*NET. Sannan CAP(t), kalmar iyawa, an kiyasta ta ne daga littattafan ma'aunin AI da rahotannin girkawa a fagen aiki. Don haka muhimmancin aiki yana fitowa ne daga daidaitaccen rumbun bayanan sana'o'i, yayin da iyawa ke fitowa daga majiyoyi biyu: gwaje-gwajen ma'auni da rahotannin filin aiki.

Takardar ta fito fili game da hanyar da ta bi, kuma mai karatu ya kamata ya rike ta a kan hakan. Ana lissafa makin ATE ne ta hanyar algorithm daga bayanan ayyuka na O*NET ta amfani da daidaitattun sigogin karbuwa; ba kiyasin regression ba ne. Hakan yana da sakamako na zahiri. Kiyasin regression yana zuwa da kayan aiki na yau da kullun na kurakuran ma'auni (standard errors) da hujjojin tantancewa. Hadadden lissafin algorithm da aka daidaita ba shi da wannan, kuma sakamakonsa hasashe ne da aka samo daga samfur (model-derived projections), kalmar da ita kanta takardar ke amfani da ita don alkaluman maye gurbin ma'aikata. Suna bayyana abin da samfurin ke nufi a karkashin sigoginsa. Ba sa ba da rahoton abin da ya riga ya faru.

Zan iya cewa wannan tsari ne na gaskiya, kuma shi ne ya dace ga matsalar hasashe inda abin da ake hasashen bai riga ya faru ba tukuna. Ma'auni mai hade da abubuwa da aka gina daga bayyanannun bayanai ana iya duba shi, a kalubalance shi kuma a sake daidaita shi. Idan wani bai yarda ba game da saurin karbuwa, ya san wace maɓalli zai motsa.

Inda samfurin ya yi amfani da shi #

Aikin gwajin yana da fadi. Takardar ta yi amfani da tsarin ATE a kan lambobin Standard Occupational Classification (SOC) 236 da suka mamaye manyan rukunin sana'o'i shida, tana samar da hasashen maye gurbin ma'aikata da aka samo daga samfur a manyan yankunan fasaha biyar na Amurka don lokacin 2025-2030. Yankunan biyar su ne San Francisco Bay Area, Seattle-Tacoma-Bellevue, Austin-Round Rock, New York-Newark, da Boston-Cambridge.

Takardar ta bayyana wannan a matsayin aikin farko na gwaji mai yankuna da yawa na tsarin fuskantar ayyuka wanda ke tunkarar tsarin agentic AI musamman. Wannan ikirari ne na kasancewa na farko da marubutan suka yi, kuma ba zan iya tantance shi daga bayanan da ke gabana ba. Abin da zan iya cewa shi ne tsarin yankunan martani ne mai ma'ana ga matsalar da binciken matakin kasa ke tarayya a ciki: fuskantar barazana siffa ce ta sana'o'i, amma karbuwa siffa ce ta wurare da kamfanoni.

A kan sakamakon da tushen bayanin ya bayar, masu sharhi kan bashi, alkalai da kwararrun dorewa sun kai makin ATE na 0.43-0.47. Wannan hadakar ta cancanci a sake duba ta. Mai sharhi kan bashi, alkali da kwararren dorewa ba sa kama da juna a horo, albashi ko aikin yau da kullun. Abin da suke da shi iri daya, a lissafin takardar, shi ne maki a cikin zango daya. Zan karanta hakan a matsayin shaida mai rushe duk wani rarraba ayyuka na kasala zuwa "masu aminci" da "masu hadari" bisa ga fanni ko bisa ga kirkire-kirkire da ake tsammani.

Zan kuma kauce wa karanta fiye da abin da tushen bayanin ke tallafawa a cikin wannan lambar. Bayanan da nake da su sun ba da zango ga wadannan sana'o'i uku kuma ba su fadi yadda yake kwatantawa da sauran lambobi 236 ba. Don haka lambar tana nuna cewa wadannan ayyukan suna fuskantar barazana mai yawa a ma'aunin takardar. Ba ta nuna cewa su ne suka fi fuskantar barazanar ba.

Abubuwan da aka gano a yankunan suna zuwa da sharadin taka-tsantsan da ya kamata a maimaita. Takardar ta bayyana cewa tasirin ma'aikata na Seattle a 2026-2027 yana ba darigar abin da ma'aikatan kasa baki daya wadanda ba na fasaha ba za su fuskanta a 2028-2030. Wannan hasashe ne mai dauke da bayyanannen jinkiri (lag), kuma an fada shi cikin taka-tsantsan. Ana karanta shi a matsayin zaton da samfurin ke samarwa, kuma bai kamata a dauke shi a matsayin tabbataccen al'amari ba. Amma nau'in ikirari ne mai amfani, domin yana da takamaiman bayani sosai har zai iya zama ba daidai ba. Idan tsarin farko na Seattle bai wuce zuwa ga ma'aikatan kasa wadanda ba na fasaha ba zuwa 2028, samfurin yana da matsalar da ba zai iya kare kansa daga gare ta ba.

Gefen dawo da sabbin ayyuka na kundin #

Makin maye gurbin ma'aikata shi kadai yana fadar rabin labarin Acemoglu-Restrepo ne kawai. A cikin tsarinsu, kason kwadago ya dogara ne akan ko kirkirar sabbin ayyuka ta fi saurin maye gurbin tsofaffi. Takardar da ta fadada tsarin kuma ta auna maye gurbi kadai za ta kasance ta watsar da bangaren da ke ba wa tsarin karfinsa.

Takardar ba ta yi hakan ba. Ta gano rukunin sana'o'i inda agentic AI ke haifar da tasirin dawo da ayyuka (reinstatement effects): sabbin ayyukan da ba su wanzu ba kafin girka tsarin agentic ya haifar da bukatar sa idon dan Adam, shugabanci da hadin gwiwa a babban ma'auni. Gaba daya ta gano rukunin sana'o'i goma sha bakwai masu tasowa wadanda ke amfana daga tasirin dawo da ayyuka. Sun tattara ne a cikin hadin gwiwar mutum da AI, shugabancin AI da ayyukan AI na musamman a fannoni daban-daban.

Tsarin yana da saukin fahimta. Idan agents suna gudanar da tsarin aiki daga farko zuwa karshe, dole ne wani ya sa ido a kansu, ya kafa dokokin da suke gudana a karkashinsu, kuma ya yi aiki kafada da kafada da su. Wadannan bukatun ba su wanzu a matsayin ayyukan yi ba a baya domin babu wani abu tukuna da za a sa wa ido.

Tushen bayanin bai gaya mini mutane nawa wadannan rukunoni goma sha bakwai za su dauka aiki ba, ko kuma ko girmansu zai cike gurbin maye gurbin da makin ATE ke nunawa. Don haka bayanan da zan iya amfani da su sun bar babbar tambayar a bude: ko dawo da sabbin ayyuka ya fi saurin maye gurbin tsofaffi. Duk wanda ya rubuta cikin karfin gwiwa cewa yana yi, ko ba ya yi, yana wuce gona da iri kan ikirarin da aka tabbatar. Gudunmawar takardar ita ce sanya dukkan bangarorin biyu na kundin a cikin tsari guda da kuma fadar sunayen inda sabbin ayyukan ke bayyana.

Abin da za a dauka daga ciki #

Akwai darasi na tsarin bincike da kuma na zahiri.

Darasin tsarin bincike shi ne na'urar bincike (unit of analysis) ita ke yanke shawarar amsar. Tsallen daga 47% zuwa 9-14% ya fito ne daga koma wa daga sana'o'i zuwa ga daidaikun ayyuka. Takardar tana cewa yanzu ana bukatar wani matakin gaba, daga daidaikun ayyuka zuwa ga daukacin tsarin aiki (workflows), domin abin da ake sarrafawa da kansa ba aiki daya ba ne yanzu. A duk lokacin da na'urar binciken ta canza, kiyasin yana canzawa, kuma tarihin fannin yana nuna cewa zabin na'urar bincike ya cancanci kulawa sosai kamar zabin bayanan da ake amfani da su.

Darasin zahiri yana ga duk wanda ke shirin adadin ma'aikata, horo ko kayayyaki a cikin shekaru biyar masu zuwa. Makin da ke gaya maka cewa kashi 80% na ayyukan wani matsayi suna fuskantar barazana bai isa ba, domin ba ya fadar ko sauran kashi 20% din jigo ne mai karko na dan Adam ko kuma dankon hadawa ne kawai wanda sashen tsare-tsare na agent zai iya samarwa yanzu. Abin da ke rufe tsarin aiki (workflow coverage factor) ya wanzu ne don yin wannan tambayar kai tsaye.

Makin ATE ma'auni ne na hadin gwiwa mai dauke da daidaitattun sigogi, kuma hasashensa an samo su ne daga samfur. Hujja ce mai tsari game da 2025-2030, ba rikodin abin da ya riga ya faru ba. Wannan dalili ne na gwada su maimakon yin biris da su. Zaton Seattle yana ba da gwaji na farko: idan tasirin 2026-2027 a can ya yi kama da yadda samfurin ke nunawa, jinkirin zuwa ga ma'aikatan kasa wadanda ba na fasaha ba a 2028-2030 zai zama abu na gaba da za a sa wa ido. Kuma rukunoni goma sha bakwai na dawo da sabbin ayyuka suna tayar da tambayar da ikirarin ba za su iya warwarewa ba: ko ayyukan da aka kirkira don sa ido kan agents za su zo da yawan da zai yi tasiri ga kason kwadago.

Tambayoyin da Ake Yawan Yi #

Me ya sa tsare-tsaren fuskantar ayyuka na baya suka raina tasirin agentic AI?

Tsare-tsaren farko da aka gina a kan al'adar Autor-Levy-Murnane suna dauka cewa sarrafa kansa yana cire daidaikun ayyuka yayin da sana'ar ke tsira a matsayin kwaryar hadin gwiwa ta dan Adam. Tsarin agentic AI yana shanye tsare-tsare, kiran kayan aiki, kula da yanayi da gwajin inganci na tsakiya wadanda a baya ke bukatar mai hadin gwiwa na mutum,ដោយ haka yana jefa daukacin tsarin aiki cikin barazana maimakon ware daidaikun ayyuka.

Ta yaya ake lissafa makin Agentic Task Exposure (ATE)?

ATE hadadden lissafi ne na algorithm da aka gina daga makin muhimmancin ayyuka na O*NET (wo,t), ma'aunin iyawar AI da rahotannin girkawa a fagen aiki (CAP(t)), workflow coverage factor da ke auna ko cikakken tsarin aikin sana'a yana shiga cikin iyakar aikin agent, da kuma daidaitattun sigogin saurin karbuwa na logistic a kan lambobin SOC 236.

Mene ne binciken ya gano game da samar da sabbin ayyuka (reinstatement effects)?

Ta hanyar fadada hanyar dawo da ayyuka ta Acemoglu-Restrepo, binciken ya gano rukunin sana'o'i goma sha bakwai masu tasowa wadanda suka tattara a cikin hadin gwiwar mutum da AI, shugabancin AI da ayyukan AI na musamman a fannoni daban-daban. Ko karuwar wadannan rukunonin za ta cike gurbin maye gurbin ma'aikata a matakin tsarin aiki ya ci gaba da kasancewa tambayar gwaji da ke bude.

Manazarta #

Ravish Gupta and Saket Kumar, (2026). Agentic AI and Occupational Displacement: A Multi-Regional Task Exposure Analysis of Emerging Labor Market Disruption ⧉. arXiv.
Tyna Eloundou, Sam Manning, Pamela Mishkin, and Daniel Rock, (2023). GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models ⧉. arXiv.
Daron Acemoglu and Pascual Restrepo, (2019). Automation and New Tasks: How Technology Displaces and Reinstates Labor ⧉. Journal of Economic Perspectives.
National Center for O*NET Development, (2025). O*NET Resource Center ⧉. US Department of Labor.

Bita ta ƙarshe 2026-11-01.

Sake buga wannan labari

Agentic AI ya rusa zaton da ke bayan kowane hasashen fuskantar ayyuka — Sebastien Rousseau

Tsare-tsaren fuskantar ayyuka suna daukar sana'a a matsayin kwarya da ke tsira a kewayen duk abin da mutane ke ci gaba da yi. Tsarin agentic yana bin manufa kuma yana isar da cikakken sakamakon aiki, don haka aikin hadin gwiwa...

An lasisin wannan labari a karkashin Creative Commons Attribution 4.0 International. Sake bugawa na bukatar nuna asalin URL na asali.

Agentic AI ya rusa zaton da ke bayan kowane hasashen fuskantar ayyuka — Sebastien Rousseau

Tsare-tsaren fuskantar ayyuka suna daukar sana'a a matsayin kwarya da ke tsira a kewayen duk abin da mutane ke ci gaba da yi. Tsarin agentic yana bin manufa kuma yana isar da cikakken sakamakon aiki, don haka aikin hadin gwiwa...

Originally published at https://sebastienrousseau.com/ha/2026-10-25-agentic-ai-rusa-zaton-ke-bayan/ by Sebastien Rousseau.
Licensed under CC-BY-4.0.

SEBASTIEN ROUSSEAU FOUNDER · ENGINEER