6.6.2 Ukuqonda nokuphatha ubungozi lwati

Ingozi yolwazi ingozi enkulu kunazo zonke ekucwaningweni kwezenhlalakahle; liye landa kakhulu; futhi kuyingozi kakhulu ukuyiqonda.

Inselele yesibili yokuziphatha ngokucwaninga nge-digital-age ingozi yolwazi , okungenzeka kube nomonakalo ekuvezweni kolwazi (National Research Council 2014) . Ulwazi olulimazayo ekudaluleni kolwazi lomuntu siqu lungaba ngumnotho (isib. Ukulahlekelwa umsebenzi), ezenhlalakahle (isb., Ukuhlazeka), kwengqondo (isib. Ukucindezeleka), noma ukuhlukumeza (isib. Ukuboshwa ngokuziphatha okungekho emthethweni). Ngeshwa, iminyaka yobudijithali ikhulisa ingozi yolwazi ngokuphawulekayo-kunolwazi oluthe xaxa mayelana nokuziphatha kwethu. Futhi ingozi yolwazi iboniswe kanzima kakhulu ukuqonda nokuphatha uma kuqhathaniswa nezingozi ezazinenkinga yokucwaninga kwezenhlalakahle ze-analog, njengengozi engokomzimba.

Enye indlela ukuthi abacwaningi social unciphise ingozi lwati kuba 'anonymization "idatha. "Anonymization" inqubo wokususa izikhombi ezisobala siqu efana negama, ikheli, kanye nenombolo yocingo kusukela idatha. Nokho, le ndlela okuningi kangako uma beqhathaniswa nabantu abaningi bayaqaphela, futhi kuba, eqinisweni, ngokujulile futhi ngokuyisisekelo limited. Ngenxa yaleso sizathu, noma nini ngichaze "anonymization," Ngiyisebenzisa bokhulunyiwe ukukukhumbuza ukuthi le nqubo kudala ukubukeka yokufihla kodwa hhayi Ukuzifihla kweqiniso.

Isibonelo esicacile sokuhluleka kwe-"anonymous" kusukela ekupheleni kweminyaka yama-1990 eMassachusetts (Sweeney 2002) . IKhomishini Yomshuwalense Yamaqembu (GIC) yayiyinhlangano kahulumeni ephethe ukuthenga umshuwalense wezempilo kubo bonke abasebenzi basezimbonini. Ngalolu msebenzi, i-GIC iqoqe amarekhodi wezempilo eningiliziwe mayelana nezinkulungwane zabasebenzi bakahulumeni. Ngomzamo wokukhuthaza ucwaningo, i-GIC yanquma ukukhulula la marekhodi kubacwaningi. Noma kunjalo, abazange bahlanganyele yonke idatha yabo; kunalokho, "babonisa" le datha ngokususa ulwazi olufana namagama namakheli. Kodwa-ke, bashiye olunye ulwazi abalucabanga ukuthi lungaba usizo kubacwaningi njengolwazi lomphakathi (i-zip, usuku lokuzalwa, ubuhlanga, nocansi) nolwazi lwezokwelapha (vakashela idatha, ukuxilongwa, inqubo) (isibalo 6.4) (Ohm 2010) . Ngeshwa, le "nonyunyiso" ayanele ngokwanele ukuvikela idatha.

Umdwebo 6.4: "Ukuqamba amanga" kuyindlela yokukhipha ulwazi olucacile lokuthola. Isibonelo, lapho ukhulula amarekhodi yomshuwalense wezokwelapha wabasebenzi base-state, i-Massachusetts Group Insurance Commission (GIC) isuse amagama namakheli kumafayela. Ngisebenzisa amanothi okucaphuna ngegama elithi "i-anonymous" ngoba inqubo inikeza ukubukeka kokungaziwa kodwa hhayi ukungaziwa kwangempela.

Ukuze sibone ukwehluleka kwe-GIC "ukuhlonishwa", uLatanya Sweeney-ke waba umfundi oqeqeshwe nge-MIT-akhokhelwa u-R20 20 ukuze athole amarekhodi okuvota edolobheni laseCambridge, idolobha laseMassachusetts umbusi uWilliam Weld. Lawa marekhodi wokuvota afaka ulwazi olufana negama, ikheli, ikhodi ye-zip, usuku lokuzalwa, nobulili. Iqiniso lokuthi ifayela ledatha yezokwelapha kanye nefayela lokuvota lakwabelana ngekhodi ye-fields-zip, usuku lokuzalwa, nocansi-lusho ukuthi i-Sweeney ingabaxhumanisa. U-Sweeney wayazi ukuthi usuku lokuzalwa lukaWeld lwangoJulayi 31, 1945, kanti amarekhodi okuvota ahlanganisa abantu abayisithupha kuphela eCambridge ngalelo suku lokuzalwa. Ngaphezu kwalokho, kulabo bantu abayisithupha, kuphela abathathu kuphela abesilisa. Futhi, kulawo madoda amathathu, ikhodi eyodwa ye-Weld eyodwa okwabelwana ngayo. Ngakho-ke, idatha yokuvota ibonise ukuthi noma ubani owedatha yezokwelapha ngokuhlanganiswa kukaWeld ngosuku lokuzalwa, ubulili, nekhodi ye-zip kwakunguWilliam Weld. Empeleni, lezi zicucu ezintathu zolwazi zanikeza iminwe yeminwe eyingqayizivele kuye kwedatha. Esebenzisa leli qiniso, u-Sweeney wakwazi ukuthola amarekhodi wezokwelapha kaWeld, futhi, ukumtshela ngaye, wamthumelela ikhophi yamarekhodi akhe (Ohm 2010) .

Umdwebo 6.5: Ukubuyiswa kabusha kwemininingwane "ye-anonymous". ULatanya Sweeney wahlanganisa amarekhodi wezempilo "angabonakali" ngamarekhodi okuvota ukuze athole amarekhodi wezokwelapha kaHulumeni uWilliam Weld Adapted kusuka ku- Sweeney (2002) , isibalo 1.

Umsebenzi ka-Sweeney ubonisa isakhiwo esiyisisekelo sokuhlaselwa kabusha kokukhomba -ukuthatha isikhathi esivela emphakathini wokuphepha kwekhompyutha. Kulezi zihlaselo, amaqoqo amabili wedatha, futhi akekho ngokwayo aveza ulwazi olubucayi, axhumene, futhi ngalokhu kuhlanganiswa, ulwazi olubucayi luvezwa.

Ekuphenduleni umsebenzi ka-Sweeney, nomunye umsebenzi ohlobene nawo, abacwaningi manje basusa ulwazi oluthe xaxa-konke okubizwa ngokuthi "ulwazi lokukhomba umuntu siqu" (PII) (Narayanan and Shmatikov 2010) -kuqhubeka nenqubo "ye-anonymous". Ngaphezu kwalokho, abacwaningi abaningi manje uqaphele ukuthi idatha ethile-njengamarekhodi wezokwelapha, amarekhodi wezezimali, izimpendulo zemibuzo yokuhlola ngokuziphatha okungekho emthethweni-mhlawumbe inokukhathazeka kakhulu ukukhulula ngisho nangemva kokuthi "i-anonymous". Nokho, izibonelo engizozozinikeza zibonisa ukuthi abacwaningi bezenhlalo badinga ukushintsha ukucabanga kwabo. Njengesinyathelo sokuqala, kuwukuhlakanipha ukuphetha ngokuthi yonke idatha engase obubonakalayo kanye yonke idatha engase ibe bucayi ukuphi. Ngamanye amazwi, kunokuba sicabange ukuthi ingozi yolwazi isebenza ngaphansi kwephrojekthi encane yamaphrojekthi, kufanele sicabange ukuthi isebenza-ngezinga elithile-kuwo wonke amaphrojekthi.

Zombili izici zalokhu kuvuselelwa kuboniswa umklomelo we-Netflix. Njengoba kuchazwe esahlukweni 5, i-Netflix ikhishwe izilinganiso ze-movie eziyizigidi ezingu-100 ezihlinzekwa ngamalunga angaba ngu-500,000, futhi zinekholi evulekile lapho abantu bevela emhlabeni wonke behambisa ukulungiswa kwamandla okungathuthukisa ikhono le-Netflix lokuncoma ama-movie. Ngaphambi kokukhulula idatha, i-Netflix isuse noma yiluphi ulwazi olubonakalayo olubonakalayo, njengamagama. Baphinde bathatha isinyathelo esengeziwe futhi bafaka ukuphazamiseka okuncane kwamanye amarekhodi (isb., Ukushintsha ezinye izilinganiso kusuka izinkanyezi ezingu-4 kuya ezinkanyezini ezingu-3). Ngokushesha baqaphela ukuthi naphezu kwemizamo yabo, idatha ayisayiziwa.

Kwaphela amasonto amabili ngemuva kokukhishwa kwedatha, u-Arvind Narayanan noVitaly Shmatikov (2008) babonisa ukuthi kungenzeka ukuthi ufunde mayelana nezintandokazi ze-movie zabantu. Ubuqili ekuhlaselweni kwabo okuhlonza kabusha kufana no-Sweeney's: hlanganisa ndawonye imithombo emibili yolwazi, enye enolwazi olubucayi futhi alukho ulwazi olucacile nolwazi oluqukethe ulwazi lwabantu. Ngamunye wale mithombo yedatha ingahle iphephile, kodwa uma ihlangene, i-dasaset ehlanganisiwe ingadala ingozi yolwazi. Endabeni ye-Netflix idatha, nansi ukuthi kungenzeka kanjani. Cabanga ukuthi ngikhetha ukwabelana ngemicabango yami mayelana nezenzo kanye nama-comedy movie nabasebenzi bami, kodwa ukuthi ngithanda ukungabelani ngombono wami mayelana namafilimu enkolo nezombangazwe. Abasebenza nabo bangasebenzisa ulwazi engabelane nabo ukuthola amarekhodi ami ku-Netflix idatha; ulwazi engingabelana ngalo lungaba yiminwe eyingqayizivele efana nohlobo luka-William Weld lokuzalwa, i-zip code, kanye nobulili. Bese-ke, uma bethola iminwe yami yeminwe eyingqayizivele yedatha, bangafunda izilinganiso zami ngazo zonke ama-movie, kufaka phakathi ama-movie engiwakhetha ukungabelani ngawo. Ngaphezu kwalolu hlobo lokuhlaselwa okuhlosiwe kugxile kumuntu oyedwa, uNayayanan no-Shmatikov babonise nokuthi kungenzeka ukuhlasela okubanzi- okubandakanya abantu abaningi-ngokuhlanganisa idatha ye-Netflix nedatha yokulinganisa yomuntu siqu kanye neyemidlalo abanye abantu abayikhethile ukuthumela kwi-Movie Movie Database (IMDb). Ngokumane nje, noma yiluphi ulwazi oluyiminwe yeminwe eyingqayizivele kumuntu othize-ngisho nesethi yabo yokulinganiswa kwe-movie-ingasetshenziswa ukuwabona.

Ngisho noma idatha ye-Netflix ingabuyekezwa kabusha kunoma yikuphi ukuhlaselwa okuhlosiwe noma okubanzi, kungase kubonakale kuyingozi kakhulu. Ngemuva kwakho konke, izilinganiso ze-movie azibonakali zibucayi kakhulu. Nakuba lokho kungase kwenzeke ngokujwayelekile, kwabanye abantu abangu-500,000 kudathasethi, izilinganiso ze-movie zingase zizwele kakhulu. Eqinisweni, ekuphenduleni ukukhonjwa kabusha, owesifazane osondelene naye wesifazane ohlangene naye ujoyine isitifiketi sokusebenza ngokumelene ne-Netflix. Nansi indlela inkinga evezwe ngayo enkantolo yabo (Singel 2009) :

"I-M] ovie nedatha yokulinganisa iqukethe ulwazi lwe ... uhlobo lomuntu siqu nobucayi kakhulu. Idatha ye-movie yelungu idalule intshisekelo yomuntu siqu kaNetflix kanye / noma ihlushwa ngezinkinga ezihlukahlukene zomuntu siqu, kubandakanya ukuziphatha ngokocansi, ukugula kwengqondo, ukuphumula kokudakwa ngokweqile, nokuhlukunyezwa okuvela emndenini, ukuhlukunyezwa ngokomzimba, ubudlova basekhaya, ukuphinga nokudlwengula. "

Ukubonakala kabusha kwedatha ye-Netflix Umklomelo kubonisa kokubili ukuthi yonke idatha ingabonakala futhi yonke idatha ingase izwele. Kuleli phuzu, ungase ucabange ukuthi lokhu kusebenza kuphela kumininingwane ekhuluma ukuthi iyabantu. Okumangalisa ukuthi akusilo iqiniso. Ukuphendula isicelo soMthetho we-Freedom of Information, uHulumeni waseNew York City wakhipha amarekhodi azo zonke izinkampani zamatekisi eNew York ngonyaka ka-2013, kufaka phakathi i-pickup futhi ilahle izikhathi, izindawo, kanye nezindleko zokuhamba (khumbula isahluko 2 ukuthi Farber (2015) basebenzisa idatha efanayo ukuhlola imibono ebalulekile embonini yezomnotho). Le datha mayelana nokuhamba kwamatekisi ingase ibonakale ingenangqondo ngoba ingabonakali ukuhlinzeka ngolwazi ngabantu, kodwa u-Anthony Tockar waqaphela ukuthi le dataset yamatekisi empeleni yayinolwazi oluningi olubucayi ngabantu. Ngokwesibonelo, wabheka zonke izinyawo eziqala kwi-Hustler Club-ibhola elikhulu le-strip eNew York-phakathi kwamabili no-6 ekuseni bese bethola izindawo zabo zokuyeka. Lokhu kusesha kwembulwe-ngokuyinhloko uhlu lwamakheli abanye abantu abavakashela i-Hustler Club (Tockar 2014) . Kunzima ukucabanga ukuthi uhulumeni womuzi wawukhumbule lokhu lapho ekhishwa idatha. Empeleni, le nqubo ingasetshenziswa ukuthola amakheli asekhaya abantu abavakashela noma iyiphi indawo edolobheni-umtholampilo wezokwelapha, isakhiwo sikahulumeni, noma isikhungo senkolo.

Lezi zimo ezimbili zeNetflix Prize kanye nedatha yeTekisi City itekisi ibonisa ukuthi abantu abanolwazi oluncane bangakwazi ukulinganisa kahle ingozi yolwazi kwimininingwane abayikhululayo-futhi lezi zimo aziyona eyingqayizivele (Barbaro and Zeller 2006; Zimmer 2010; Narayanan, Huey, and Felten 2016) . Ngaphezu kwalokho, ezimweni eziningi ezinjalo, idatha enenkinga isatholakala ngokukhululekile ku-intanethi, ebonisa ubunzima bokuhlehlisa ukukhishwa kwedatha. Ngokubambisana, lezi zibonelo-kanye nokucwaninga kwikhompyutha yesayensi mayelana nobumfihlo-kuholela esiphethweni esibalulekile. Abacwaningi okufanele acabange ukuthi yonke idatha ezingaba obubonakalayo kanye yonke idatha engase ibe bucayi ukuphi.

Ngeshwa, asikho isixazululo esilula emaqinisweni ukuthi yonke idatha ingabonakala futhi yonke idatha ingase izwele. Kodwa-ke, enye indlela yokunciphisa ingozi yolwazi ngenkathi usebenza nedatha ukudala nokulandela uhlelo lokuvikelwa kwedatha . Lolu hlelo luzokwehlisa ithuba lokuthi idatha yakho izovuza futhi izokwehlisa umonakalo uma kwenzeka ukuvuza okuthile ngandlela-thile. Ukucaciswa kwezinhlelo zokuvikela idatha, njengokuthi yikuphi uhlobo lokubethelwa kwedatha okuzosetshenziswa, kuzoshintsha ngokuhamba kwesikhathi, kodwa i-UK Data Services ihlela ngokuhlelekile izakhi zepulani lokuvikela idatha ngezigaba ezinhlanu ezibiza ngokuthi izintambo ezinhlanu : amaphrojekthi ephephile, abantu abaphephile , izilungiselelo eziphephile, idatha ephephile, kanye nemiphumela ephephile (ithebula 6.2) (Desai, Ritchie, and Welpton 2016) . Akekho kulezi ziphephe ezinhlanu ezihlinzeka ngokuvikelwa okuphelele. Kodwa ndawonye benza isethi esinamandla sezinto ezinganciphisa ingozi yolwazi.

Ngaphezu kokuvikela idatha yakho ngenkathi usebenzisa, isinyathelo esisodwa kwinqubo yokucwaninga lapho ingozi yolwazi ikakhulukazi ingabelana khona ukwabelana kwedatha nabanye abacwaningi. Ukukwabelana kwedatha phakathi kososayensi kuyigugu eliyinhloko lomzamo wesayense, futhi kusiza kakhulu ukuthuthukiswa kolwazi. Nakhu ukuthi i-UK House of Commons yichaza ukubaluleka kokwabelana kwedatha (Molloy 2011) :

"Ukufinyeleleka kwedatha kuyisisekelo uma abacwaningi bezokhiqiza, baqinisekise futhi bakhiwe kwimiphumela ebikwe ezincwadini. Ukucatshangelwa kufanele kube ukuthi, ngaphandle uma kunesizathu esinamandla ngenye indlela, idatha kufanele ivezwe ngokugcwele futhi yenziwe obala. "

Noma kunjalo, ngokwabelana ngedatha yakho nomunye umcwaningi, ungase ukwandise ingozi yolwazi kubahlanganyeli bakho. Ngakho-ke, kungase kubonakale ukuthi ukwabelana kwedatha kubangela ukungezwani okuyisisekelo phakathi kwesibopho sokukwabelana ngolwazi nabanye ososayensi kanye nesibopho sokunciphisa ingozi yolwazi kubahlanganyeli. Ngenhlanhla, le nkinga ayinzima kakhulu njengoba ibonakala. Kunalokho, kungcono ukucabanga mayelana nokwabelana kwedatha njengoba kuwela ngokuqhubekayo, nephuzu ngalinye kulokho kuqhubekela phambili okuhlinzeka ngokuxuba okuhlukile kwezinzuzo emphakathini kanye nobungozi kubahlanganyeli (isibalo 6.6).

Ngokweqile, ungabelana ngedatha yakho ngaphandle komuntu, okunciphisa ingozi kubahlanganyeli kodwa futhi kunciphisa inzuzo emphakathini. Ngakolunye uhlangothi, ungakhulula futhi ukhohlwe , lapho idatha "ingabonakali" futhi ithunyelwe kuwo wonke umuntu. Ngokuphathelene nokukhipha idatha, ukukhululwa nokukhohlwa okunikezwayo kokubili izinzuzo eziphakeme emphakathini kanye nengozi ephezulu kubahlanganyeli. Phakathi kwalezi zimo ezimbili ezimbi kakhulu kukhona uhlobo lwezinhlamvu, kufaka phakathi ukuthi ngizobiza kanjani indlela engadini ekhonjiwe . Ngaphansi kwalendlela, idatha ikwabelwana nabantu abahlangabezana nemigomo ethile futhi abavuma ukuboshwa ngemithetho ethile (isb., Ukubhekwa kwe-IRB kanye nohlelo lokuvikelwa kwedatha). Indlela yokuvotela engadini ihlinzeka ngezinzuzo eziningi zokukhululwa futhi ukhohlwe ngengozi encane. Yiqiniso, indlela enjalo yenza imibuzo eminingi-obani okufanele bafinyelele, ngaphansi kweziphi izimo, nokuthi bangakanani okufanele bakhokhe ukuze balondoloze futhi baphathe ipulazi elinogada, njll-kodwa lokhu akunakwenzeka. Eqinisweni, sekukhona kakade izivande ezinezivalo ezikhona lapho abacwaningi abangasebenzisa khona njengamanje, njengengqolobane yedatha ye-Inter-University Consortium yezoPolitiki nezeNhlalakahle eNyuvesi yaseMichigan.

Umdwebo 6.6: Amasu okukhululwa kwedatha angawela ngokuqhubekayo. Lapho kufanele ube khona kulokhu kuqhubeka kuxhomeke emininingwaneni ethile yedatha yakho, futhi ukubuyekezwa komuntu wesithathu kungakusiza ukunquma ibhalansi efanele futhi uzuze esimweni sakho. Ukuma ngokuqondile kwalesi sigqoko kuncike emibhalweni yemininingwane kanye nemigomo yocwaningo (Goroff 2015) .

Ngakho-ke, idatha kufanele isuke kuphi ekutadisheni kwakho ekuhambeni kokungahlanganyeli, ingadi enamatshe, nokukhululwa futhi ukhohlwe? Lokhu kuxhomeke emininingwaneni yedatha yakho: abacwaningi kumele balinganise Ukuhlonipha Abantu, Ukuzuza, Ubulungiswa, Nokuhlonipha Umthetho Nesithakazelo Somphakathi. Kubhekwe kulokhu okubheka, ukukwabelana kwedatha akuyona i-conundrum yokuziphatha ehlukile; kuyisinye sezici eziningi zocwaningo lapho abacwaningi kudingeka bathole ibhalansi efanelekayo yokuziphatha.

Abanye abagxeki ngokuvamile baphikisana nokwabelana kwedatha ngoba, ngokombono wami, bagxile ezingozini zayo-okungabazeki ukuthi bangempela-futhi bayayinaki izinzuzo zayo. Ngakho-ke, ukuze sikhuthaze ukugxila kokubili izingozi nezinzuzo, ngingathanda ukunikeza ukufana. Njalo ngonyaka, izimoto zibhekene nezinkulungwane zabantu abafa, kodwa asizami ukuvimbela ukushayela. Empeleni, ukushayela ukuvinjelwa kungavamile ngoba ukushayela kwenza izinto eziningi ezinhle. Esikhundleni salokho, umphakathi ubeka imingcele kulabo abangakwazi ukushayela (isib. Isidingo sokuba nesikhathi esithile nokuthi baqede ukuhlolwa okuthile) nokuthi bangakwazi kanjani ukushayela (isib., Ngaphansi komkhawulo wejubane). Umphakathi unabantu abanomsebenzi wokuphoqelela le mithetho (isib. Amaphoyisa), futhi sijezisa abantu abanjwe abahlukumezayo. Loluhlobo olufanayo lokucabanga okulinganiselayo umphakathi olusebenzayo ekulawuleni ukushayela lungasetshenziswa futhi ekukwabelaneni kwedatha. Okungukuthi, kunokuba senze izimpikiswano eziphelele noma ngokumelene nokwabelana kwedatha, ngicabanga ukuthi sizokwenza intuthuko enkulu ngokugxila endleleni esingayinciphisa ngayo izingozi bese sandisa izinzuzo kusuka kokwabelana kwedatha.

Ukuphetha, ingozi yolwazi yande kakhulu, futhi kunzima ukubikezela nokulinganisa. Ngakho-ke, kungcono ukuthi ucabange ukuthi yonke idatha ingabonakala futhi ingase izwele. Ukwehlisa ingozi yolwazi ngenkathi ucwaninga, abacwaningi bangadala futhi balandele uhlelo lokuvikelwa kwedatha. Ngaphezu kwalokho, ingozi yolwazi ayivimbeli abacwaningi ukuba bahlanganyele idatha nabanye ososayensi.