Ama-algorithms we-Franken ahamba kabi

ISIKWELETU SESITHOMBE:

iStock

I-Franken-Algorithms: Ama-algorithms ahambile

Ngokuthuthuka kobuhlakani bokwenziwa, ama-algorithms avela ngokushesha kunalokho obekulindelwe abantu.

About the Author:
Igama lomlobi
I-Quantumrun Foresight
April 12, 2023

Njengoba ukufunda komshini (ML) ama-algorithms ethuthuke kakhulu, ayakwazi ukufunda futhi azivumelanise namaphethini kumadathasethi amakhulu ngokwawo. Le nqubo, eyaziwa ngokuthi "ukufunda okuzenzakalelayo," ingaholela ekutheni i-algorithm ikhiqize eyayo ikhodi noma imithetho yokwenza izinqumo. Inkinga ngalokhu ukuthi ikhodi ekhiqizwe i-algorithm ingase ibe nzima noma ingenzeki kubantu ukuba bayiqonde, okwenza kube inselele ukukhomba ukuchema.

Umongo we-Franken-Algorithms

I-Franken-Algorithms ibhekisela kuma-algorithms (imithetho elandelwa amakhompuyutha lapho ecubungula idatha futhi ephendula imiyalo) eseyinkimbinkimbi futhi yahlangana kangangokuthi abantu abakwazi ukuyichaza. Leli gama lihambisana nenganekwane yesayensi kaMary Shelley mayelana "nesilwane esisabekayo" esadalwa usosayensi ohlanya uDkt. Frankenstein. Nakuba ama-algorithms namakhodi kuyizingqimba zokwakha zobuchwepheshe obukhulu futhi avumele i-Facebook ne-Google ukuthi kube izinkampani ezinethonya eziyizo manje, kusekuningi kakhulu ngobuchwepheshe abantu abangabazi.

Lapho abahleli bezinhlelo bakha amakhodi futhi bawasebenzise ngesofthiwe, i-ML ivumela amakhompyutha ukuthi aqonde futhi abikezele amaphethini. Nakuba ubuchwepheshe obukhulu buthi ama-algorithms anenhloso ngoba imizwa yabantu nokungabikezeli akubathinti, lawa ma-algorithms angashintsha futhi abhale imithetho yawo, okuholela emiphumeleni eyinhlekelele. Ikhodi ekhiqizwa yilawa ma-algorithms ngokuvamile iyinkimbinkimbi futhi ingabonakali, okwenza kube nzima kubacwaningi noma ochwepheshe ukuhumusha izinqumo ze-algorithm noma ukukhomba noma yikuphi ukuchema okungase kube khona kunqubo yokwenza izinqumo ye-algorithm. Lesi sivimbelo singadala izinselele ezibalulekile kumabhizinisi athembele kulawa ma-algorithms ukwenza izinqumo, njengoba engase angakwazi ukuqonda noma ukuchaza ukucabanga ngalezo zinqumo.

Umthelela ophazamisayo

Lapho ama-Franken-Algorithms ehamba kabi, kungaba yindaba yokuphila nokufa. Isibonelo kwaba yingozi yango-2018 lapho imoto ezishayelayo e-Arizona ishayisa yabulala owesifazane owayegibele ibhayisikili. Ama-algorithms emoto awakwazanga ukumkhomba ngendlela efanele njengomuntu. Ochwepheshe badabukile ngembangela yengozi—ingabe imoto yayihlelwe ngendlela engafanele, futhi ingabe i-algorithm yaba yinkimbinkimbi kakhulu ukuze izuze yona? Lokho abahleli bohlelo abangavumelana ngakho, nokho, ukuthi kudingeka kube nohlelo lokwengamela lwezinkampani zama-software—ikhodi yezimiso zokuziphatha.

Kodwa-ke, le khodi yezimiso zokuziphatha iza nokuhlehla okuthile okuvela kubuchwepheshe obukhulu ngoba basebhizinisini lokuthengisa idatha nama-algorithms, futhi abakwazi ukukhokhela ukulawulwa noma kudingeke ukuthi babonise ngale. Ukwengeza, intuthuko yakamuva ebangele ukukhathazeka kubasebenzi abakhulu bezobuchwepheshe ukwanda kokusetshenziswa kwama-algorithms ngaphakathi kwezempi, njengokubambisana kwe-Google noMnyango Wezokuvikela wase-US ukuze kufakwe ama-algorithms kubuchwepheshe bezempi, njengama-autonomous drones. Lolu hlelo lokusebenza luholele ekutheni abanye abasebenzi bashiye phansi futhi ochwepheshe bazwakalise ukukhathazeka ngokuthi ama-algorithms awakaqageleki kakhulu ukuthi angasetshenziswa njengemishini yokubulala.

Okunye okukhathazayo ukuthi i-Franken-Algorithms ingase iqhubekisele phambili futhi ikhulise ukuchema ngenxa yamadathasethi aqeqeshwe kuwo. Le nqubo ingaholela ezindabeni ezihlukahlukene zomphakathi, okuhlanganisa ukucwasa, ukungalingani, nokuboshwa ngokungafanele. Ngenxa yalezi zingozi eziphakeme, izinkampani eziningi zobuchwepheshe seziqala ukushicilela imihlahlandlela yazo ye-AI yokuziphatha ukuze icace endleleni ezithuthukisa ngayo, zisebenzisa, futhi ziqaphe ama-algorithms azo.

Imithelela ebanzi ye-Franken-Algorithms

Imithelela engaba khona ye-Franken-Algorithms ingafaka:

Ukuthuthukiswa kwezinhlelo ezizimele ezikwazi ukwenza izinqumo futhi zithathe izinyathelo ngaphandle kokuqondisa komuntu, okuphakamisa ukukhathazeka mayelana nokuziphendulela nokuphepha. Kodwa-ke, ama-algorithms anjalo anganciphisa izindleko zokuthuthukisa isoftware namarobhothi angakwazi ukwenza umsebenzi womuntu ngokuzenzakalelayo kuzo zonke izimboni.
Ukucutshungulwa okwengeziwe kokuthi ama-algorithms angenza kanjani ubuchwepheshe bezempi ngokuzenzakalelayo futhi asekele izikhali ezizimele nezimoto.
Ingcindezi eyengeziwe yokuthi ohulumeni nabaholi bezimboni basebenzise ikhodi ye-algorithm yezimiso zokuziphatha nezimiso.
Ama-Franken-Algorithms athinta ngokungafanele amaqembu athile abantu, njengemiphakathi ehola kancane noma isibalo sabantu abambalwa.
I-Franken-Algorithms ingase iqhubekisele phambili futhi ikhulise ukucwasa nokuchema ekuthathweni kwezinqumo, ezifana nezinqumo zokuqasha nokuboleka.
Lawa ma-algorithms asetshenziswa izigebengu ze-inthanethi ukuze ziqaphe futhi zisebenzise ubuthakathaka ezinhlelweni, ikakhulukazi ezikhungweni zezimali.
Abalingisi bezepolitiki abasebenzisa ama-algorithms akhohlakele ukwenza imikhankaso yokumaketha ngokuzenzakalela besebenzisa amasistimu e-AI akhiqizayo ngezindlela ezingathonya umbono womphakathi futhi ziguqule ukhetho.