6.6.2 Ukuqonda nokulawula umngcipheko zolwazi

Umngcipheko ulwazi umngcipheko ixhaphakileyo kuphando kwezentlalo; ngoku izazekile kakhulu; kwaye ubungozi kunzima ukuqonda.

Umngeni wesibini ngokweenqobo kukwaluphala lwentlalo uphando lwedijithali umngcipheko ingcaciso, amandla senzane ukusuka ukuxelwa kolwazi (Council 2014) . sidale ingozi yolwazi evela ukuxelwa kolwazi yobuqu inokuba kwezoqoqosho (umzekelo, ukuphelelwa ngumsebenzi), kwezentlalo (umzekelo, ukuphoxeka), ngokwasengqondweni (umzekelo, ukudandatheka), okanye mthetho (umzekelo, ukubanjwa ngenxa ukuziphatha ngokungekho). Ngelishwa, ubudala yedijithali kwandisa kakhulu-kukho ulwazi umngcipheko nje kangaka inkcazelo engakumbi malunga ukuziphatha kwethu. Kwaye, umngcipheko ngolwazi kubonakele kunzima kakhulu ukuqonda nokulawula kuthelekiswa ubungozi ezibe iinkxalabo kuphando zentlalo analog ubudala, ezifana nengozi emzimbeni. Ukuze ubone indlela ubudala yedijithali kwandisa umngcipheko zolwazi, cinga utshintsho ukusuka iphepha kwiirekhodi zonyango elektroniki. Zombini iintlobo zeerekodi kudala umngcipheko, kodwa iirekhodi electronic idale ingozi enkulu kakhulu ngenxa yokuba isikali inkitha no ku sulela iqela ongagunyaziswanga okanye ladityaniswa nezinye iirekhodi. abaphandi loLuntu kweminyaka yedijithali sele ndabaleka abe sengxakini umngcipheko zolwazi, ngokuyinxenye ngenxa yokuba akazange aqonde ngokupheleleyo ukuba ubungakanani nokulawula njani. Ngoko ke, ndiza ukunika indlela luncedo ukuba bacinge umngcipheko zolwazi, yaye emva koko ndiza kukunika ezinye iingcebiso nendlela ukulawula umngcipheko yolwazi kuphando wakho ukukhupha iinkcukacha kwezinye abaphandi.

Enye indlela, ukuba abaphandi lwentlalo ukunciphisa umngcipheko zolwazi yi "anonymization" ye data. "Anonymization" yinkqubo yokususa izinto zobuqu ezicacileyo njengegama, idilesi, nenombolo yomnxeba evela data. Nangona kunjalo, le ndlela kakhulu ngempumelelo kunokuba abantu abaninzi beqaphela, yaye, enyanisweni, ngokunzulu ngokuncamisayo elithile. Ngenxa yeso sizathu, xa ndithe ukuchaza "anonymization," Ndiza usebenzise iimpawu zocaphulo sinikhumbuze ukuba le nkqubo yenza imbonakalo ukungaziwa kodwa ukungaziwa oyinyaniso.

Umzekelo ocacileyo yokungakwazi "anonymization" livela 1990 kade Massachusetts (Sweeney 2002) . I-Inshorensi Commission Group (idlale) waye arhente karhulumente anoxanduva ukuthenga inshorensi yempilo kubo bonke abasebenzi bakarhulumente. Ngalo msebenzi, ukuba idlale eqokelelwe iingxelo zempilo malunga amawaka abaqeshwa bakarhulumente. Kwiinzame ku khuthaza uphando ngeendlela zokuphucula impilo, kuba idlale wagqiba ukuba ndinikhululele ezi rekhodi abaphandi. Noko ke, abazange ukwabelana yonke idatha yabo; kunoko, "zingachazwa ukuba zezikabani" oko ngokususa iinkcukacha ezifana negama nedilesi. Noko ke, abashiya ezinye iinkcukacha ukuba babecinga kuba luncedo kubaphandi ezifana nolwazi lwabantu (ikhowudi zip, umhla wokuzalwa, ubuzwe, kunye nokwabelana ngesondo) kunye nolwazi lwezonyango (data utyelelo, uxilongo, inkqubo) (Figure 6.4) (Ohm 2010) . Ngelishwa, le "anonymization" awonelanga ukukhusela data.

Isazobe 6.4: "Anonymization" yinkqubo yokususa ngokucacileyo echaza ulwazi. Ngokomzekelo, xa ukhupha iirekhodi inshorensi yonyango labasebenzi karhulumente Massachusetts Insurance Commission Group (idlale) kususwa igama nedilesi ukusuka kwifayile. Ndisebenzisa kuchaphulo engqonge igama "anonymization" kuba le nkqubo ibonelela imbonakalo ukungaziwa, kodwa ukungaziwa elililo.

Ngokomzekelo iintsilelo le idlale "anonymization", Latanya Sweeney-ke umfundi kwezidanga MIT-ihlawulwe $ 20 ukuze bafumane iirekhodi okuvota ukusuka kwisixeko yaseCambridge, idolophu Massachusetts irhuluneli William Weld. Ezi rekhodi uvotela kuquka ulwazi njengegama, idilesi, inombolo yelizwe, umhla wokuzalwa, kunye nesini. Inyaniso ukuba ifayile zonyango data kwaye ifayile lwabavoti ekwabelwana ikhowudi amasimi-zip, umhla wokuzalwa, kwaye sex-kwakuthetha ukuba Sweeney wayenokukwazi ukuzinxulumanisa kubo. Sweeney wayesazi ukuba yokuzalwa Weld yaba July 31, 1945, kwaye ingxelo okuvota ubandakanye abantu abathandathu kuphela Cambridge naloo lokuzalwa. Ngaphezu koko, abo bantu abathandathu, abathathu kuphela eyindoda. Kwaye, kuloo madoda mathathu, omnye kuphela ekwabelwana ikhowudi zip Weld ngayo. Ngoko ke, i data okuvota wabonisa ukuba nabani na iinkcukacha zonyango kunye indibanisela Weld ngayo umhla wokuzalwa, isini, kunye nekhowudi yezip yaba nguWilliam Weld. Enyanisweni, ezi amathathu ngokwengcaciso eyodwa eseminweni kuye data. Ukusebenzisa le nyaniso, Sweeney ndakwazi ukufumana iingxelo zonyango Weld, kwaye ukwazisa kuye feat wakhe, yena imeyili kuye ikopi iirekhodi zakhe (Ohm 2010) .

Isazobe 6.5: Re-idenification data "zingachazwa ukuba zezikabani". Latanya Sweeney edibanisa umthetho "zingachazwa ukuba zezikabani" iirekhodi ezempilo kunye neerekhodi ukuvota ukuze sifumane iingxelo zonyango kweRhuluneli William Weld (Sweeney 2002) .

Umsebenzi sweeney usibonisa isakhiwo esisiseko ukuhlaselwa de-anonymization -ukuba amkele elide evela kubahlali yokhuseleko computer. Kwezi ukuhlaselwa, iiseti ezimbini data, akukho nto iyodwa ibonisa ontununtunu, zinxulumana, yaye ngalo yePSO, ontununtunu zakohlwaywa. Ngandlel 'ithile le nkqubo iyafana nendlela nesoda yokubhaka kwakunye neviniga, izinto zombini bebodwa ekhuselekileyo, zinokudityaniswa ukuvelisa isiphumo elibi.

Ukusabela umsebenzi Sweeney, kunye neminye imisebenzi enxulumene, ngoku abaphandi ukususa ngokubanzi phi ulwazi-wonke ukuze ngokuba "Iinkcukacha Umntu Ikuko" (PII) (Narayanan and Shmatikov 2010) ngamaxesha inkqubo 'anonymization. "Ngaphezu koko, abaphandi abaninzi ngoku ndaqonda ukuba ezithile data-ezifana iingxelo zonyango, iirekhodi zezimali, iimpendulo kulibhala imibuzo malunga ngokungekho ukuziphatha-mhlawumbi olubuthathaka kakhulu ukuba ndinikhululele nasemva "anonymization." Noko ke, imizekelo engakumbi yakutshanje ukuba ndiza ukuchaza ngezantsi zibonisa ukuba abaphandi zoluntu kufuneka ukutshintsha indlela abacinga ngayo. Njengoko njengenyathelo lokuqala, kuya kuba bubulumko ukucinga ukuba yonke idatha ezinokuba achongeka kwaye yonke idatha ezinokuba novakalelo. Ngamanye amazwi, kunokuba ucinga ukuba umngcipheko zolwazi isebenza elucwangciso olusezantsi encinane lweeprojekthi, simele sicinge ukuba usebenza-kwezinye isidanga-zonke iiprojekthi.

Zombini ezi nkalo kule yokuqhelaniswa-ibonisiwe yi-Netflix Prize. Njengoko kuchaziwe kwiSahluko 5, Netflix wakhulula 100 lezigidi movie amanqaku anikwe ngamalungu phantse-500,000, yaye ifowuni evulelekileyo apho abantu bevela kulo lonke ihlabathi zingeniswe algorithms ukuba bakwazi ukuphucula ubuchule Netflix ukuba licebise iimovie. Phambi kokukhupha data, Netflix wazisusa naluphi na ulwazi ngokucacileyo ngokobuqu-ekuchongeni, ezifana ngamagama. Kwakhona Netflix waya inyathelo engaphezulu kwaye waqalisa perturbations okungephi kwezinye iirekhodi (umzekelo, ukutshintsha ezinye ratings kweenkwenkwezi 4 ukuya iinkwenkwezi 3). Kungekudala Netflix wafumanisa, nangona kunjalo, ukuba nangona imigudu yabo, idatha zazingathethi abangaziwayo.

Kwiiveki nje ezimbini emva data zakhutshwa Narayanan and Shmatikov (2008) babonisa ukuba kwakunokwenzeka ukuba bafunde ezithandwa ngabantu abathile xa oshukumayo. Iqhinga ukuze ihlasela kwakhona yesazisi-elifana Sweeney esithi: kuhlanganiswe ndawonye imithombo ezimbini ulwazi, omnye iinkcukacha ezinokuba ontununtunu kwaye akukho ulwazi ngokucacileyo ekuchongeni enye equlathe iinkcukacha zabantu. Nganye kwezi imithombo yedatha zisenokuba ngamnye ekhuselekileyo, kodwa xa zidibene le dataset edityanisiweyo ungenza umngcipheko zolwazi. Kwimeko ye data Netflix, nantsi oko kwenzeka njani. Cinga nje ukuba ukhethe ukwabelana iingcinga zam malunga amanyathelo kunye iimovie comedy kunye wam osebenza nabo, kodwa ukuba ndikhetha ukuba ukwabelana uluvo lwam malunga iimovie yonqulo nezopolitiko. Endisebenza nabo bangasebenzisa ulwazi Ndibalele nabo ukuba bafumane iirekhodi yam data Netflix; ulwazi ukuze ndibe lidlelane kuba esisodwa eseminweni kanye umhla William Weld ukuzalwa, ikhowudi zip, kunye nesini. Ngoko, xa befumanisa ukuba umnwe wam lelikhethekileyo data, uze bafunde amanqaku zam zonke iimovie, kubandakanywa eshukumayo apho ukhetha ukwabelana. Ukongeza koku uhlobo kuhlaselwa ekujoliswe ingqalelo kumntu omnye, Narayanan and Shmatikov (2008) babonisa ukuba kwakunokwenzeka ukwenza ebanzi ukuhlaselwa -enye ebandakanya abantu abaninzi-ngokuthi nyanisa i data Netflix idatha rating buqu kunye imovie ukuba abanye abantu bakhetha sithuba e bamanye amagamaQuery (IMDb). Naluphi ulwazi Fingerprints eyodwa kumntu-nkqu ethile iseti yabo movie ratings-ingasetyenziswa ukuchonga kubo.

Nangona data Netflix ziphinde ezichongiweyo-in nokuba ukuhlaselwa ekujoliswe okanye elinesiseko esibanzi, nto nangoku, ukuze ungabonakali ebantwini ukuba ingozi ephantsi. Ngapha koko, amanqaku bhanyabhanya ayibakhathazi ethe kakhulu. Nangona oko kunokuba yinyaniso ngokubanzi, kuba abanye 500,000 kwi dataset, ratings imovie kunokuba yinto novelwano. Enyanisweni, ekhawulelana de-anonymization umfazi lesbian luvaleleke wazibandakanya isuti udidi-amanyathelo Netflix. Nantsi indlela ingxaki yabonakaliswa isimangalo yabo (Singel 2009) :

"[M] ovie yaye ireyithingi data iqulethe iinkcukacha zohlobo buqu ngaphezulu kakhulu kwaye sensitive [sic]. movie data yelungu ayifihlayo umdla ilungu Netflix ngayo buqu kunye / okanye imizabalazo kunye ezahlukeneyo kwimiba yobuqu kakhulu, kubandakanywa isini, isigulo sengqondo, nokufunyanwa lotywala, kwaye amaxhoba nombulo, gadalala ngokwasemzimbeni, ubundlobongela basekhaya, ukukrexeza, kunye nokudlwengulwa. "

I-de-anonymization ye data Netflix Prize ubonisa zombini ukuba yonke idatha ezinokuba achongeka kwaye yonke idatha ezinokuba novakalelo. Kweli nqanaba, usenokucinga ukuba oku kusebenza kuphela data loo oluchazwa ngabantu. Okumangalisayo kukuba, oko akunjalo. Ukuphendula Inkululeko isicelo Law Ulwazi, le New York City uRhulumente wakhupha iirekhodi zonke ukhwele eNew York iteksi ngo-2013, kubandakanywa veni kwaye lokothula amaxesha, neendawo, kunye neemali yokukhwela (khumbula ku Isahluko 2 Farber (2015) wasebenzisa eli data ukuvavanya iingcamango ebalulekileyo Economics zabasebenzi). Nangona le nkcukacha malunga Iihambo iteksi kunokubonakala nabungozi kuba akafuni kubonakala ukuba ulwazi malunga nabantu, Anthony Tockar waqonda ukuba le dataset iteksi okunene eziqulathwe ezininzi iinkcukacha ezinokuba novakalelo malunga abantu. Ngokomzekelo, wakhangela zonke iihambo uqala Hustler Club-estrip club elikhulu eNew York-phakathi kobusuku, 6am ndaza ndafumana iindawo zabo drop-off. Le uphendlo kutyhilwa-ingumongo-uluhlu lweedilesi abanye abantu abadla Hustler Club (Tockar 2014) . Kunzima ukucinga ukuba urhulumente sixeko oku engqondweni xa lakhupha idatha. Enyanisweni, lo efanayo ezino kusetyenziswa ukufumana iidilesi kwikhaya labantu tyelela naliphi na indawo esixekweni-ikliniki yempilo, isakhiwo sikarhulumente, okanye iziko yonqulo.

Le meko-babini Prize Netflix kunye City New York iteksi data-zibonisa ukuba abantu abanezakhono ngokwentelekiso basilela ukuqikelela ngokuchanekileyo umngcipheko yolwazi kwi data ukuba ukhululwe, yaye ezi meko Akunguye nakanye kuthetha ezizodwa (Barbaro and Zeller Jr 2006; Zimmer 2010; Narayanan, Huey, and Felten 2016) . Ngaphezu koko, kwiindawo ezininzi kwezi meko, iinkcukacha yingxaki olusekhona ngokukhululekileyo kwikhompyutha, ebonisa ukusokola kuze okwa ukukhululeka data. Bebonke le mizekelo-kwakunye uphando kwinzululwazi yekhompyutha malunga umgaqo-kukhokelela wenza isigqibo esibalulekileyo. Abaphandi kufuneka sicinge ukuba yonke idatha ezinokuba achongeka kwaye yonke idatha ezinokuba novakalelo.

Ngelishwa, akukho isisombululo elula yokuba yonke idatha ezinokuba achongeka kwaye yonke idatha ezinokuba novakalelo. Noko ke, enye indlela zokunciphisa umngcipheko ulwazi xa usebenza data kukudala nokulandela isicwangciso ukhuseleko data. Esi sicwangciso iza kunciphisa ithuba lokuba data enokuvuza kwaye uya ukunciphisa umonakalo ukuba ukuvuza ithile kwenzeka. Ozifumene lwezicwangciso ukhuseleko data, ezifana apho uhlobo iinkropshini ukusebenzisa, aya kutshintsha ekuhambeni kwexesha, kodwa UK Iinkonzo Data zayincedayo uqulunqa iziqalelo isicwangciso ukhuseleko data ngokweendidi 5 Bayibiza iisefu 5: Iiprojekthi ekhuselekileyo, abantu ekhuselekileyo , izicwangciso ekhuselekileyo, iinkcukacha ekhuselekileyo, kunye neziphumo ekhuselekileyo (Uluhlu 6.2) (Desai, Ritchie, and Welpton 2016) . Namnye iisefu ezintlanu ngamnye ukubonelela ukhuseleko egqibeleleyo. Kodwa, kunye zenza iseti ezinamandla ezenza ukuba sikwazi ukunciphisa umngcipheko zolwazi.

Ukongeza ekukhuseleni data yakho xa usebenzisa oko, elinye inyathelo kwinkqubo yophando apho umngcipheko zolwazi ngakumbi siprofe- ixhomekeke data nabanye abaphandi. yokwabelana Data phakathi izazinzulu lixabiso core seenzame lwezenzululwazi, kwaye izibonelelo kakhulu phambili ulwazi. Nantsi indlela iNdlu UK of Commons wachaza ukubaluleka yokwabelana data:

"Ukufikelela data sisisekelo ukuba abaphandi babe lokuyivelisa, siqinisekise kwaye wakhele kwiziphumo ezixelwa kulo ncwadi. Eyona nto kufuneka ukuba, ngaphandle kokuba kukho isizathu esinamandla kungenjalo, iinkcukacha kufuneka idizwe ngokupheleleyo yaye fumaneka. Ngokuhambelana nalo mgaqo, apho kunokwenzeka, idatha enxulumene lonke uphando ezixhaswa ngemali ngurhulumente kufuneka zenziwe ngokubanzi kwaye kufumaneka ngesisa. " (Molloy 2011)

Sekunjalo, ngokwabelana data yakho komnye umphengululi, kusenokwenzeka ukuba ukwanda umngcipheko yolwazi kwi nxaxheba zakho. Ngenxa yoko, kusenokubonakala ngathi ukuba abaphandi banqwenela ukuba babelane babo data-okanye kufuneka ukuba babelane babo data-zijongene ukungaqondani esisiseko. Kwelinye icala kuba ndimelwe yeenqobo ukwabelana data zazo kunye nezinye izazinzulu, ingakumbi ukuba uphando santlandlolo ngemali ngurhulumente. Noko ke, ngelo xesha, abaphandi amele yeenqobo ukunciphisa, kangangoko kunokwenzeka, ubungozi ulwazi nxaxheba yabo.

Ngethamsanqa, le ngxaki asinto ubunzima njengoko ibonakala. Kubalulekile ukucinga data zoqhubo uhambahamba continuum evela akukho kwabelana data ukumkhulula bakulibale, apho iinkcukacha "zingachazwa ukuba zezikabani" kwaye ngeposi ukuze nabani ukufikelela (Figure 6.6). Zombini ezi zithuba ebaxiweyo kuba iingozi kunye neenzuzo. Oko kukuthi, loo nto ayithethi nto yokuziphatha ukuba ungabelani data yakho; loo ndlela kushenxisa iinzuzo ezininzi kuluntu. Ukubuyela Ukungcamla, sies, kwaye Time, umzekelo sikhe satsho apha kwisahluko, iingxoxo malunga ukukhululwa data kugxininisa kuphela ekonzakalisweni kunokwenzeka kwaye batyeshele iinzuzo kunokwenzeka bengqongqo macala onke; Ndiza ukuchaza ngxaki ngale nto, indlela enye-eve ngokugqithiseleyo olukhuselayo ngakumbi apha ngezantsi xa Ndikubekela amacebiso ukwenza izigqibo ubuso ukungaqiniseki (Icandelo 6.6.4).

Isazobe 6.6: izicwangciso ukukhululwa Data ungawa uhambahamba continuum. Apho kufuneka ecaleni kule okuqhubeka kuxhomekeka kwiinkcukacha ezithile data yakho. Kule meko, kwakhona iqela lesithathu kungakunceda isigqibo uzinzo olufanelekileyo umngcipheko nenzuzo ityala lakho.

Ngaphezu koko, phakathi kwezi meko zimbini ezigabadeleyo into endiya ebizwa indlela nesitiya esibiyelweyo apho data kwabelwana nabantu abazanelisayo iikhrayitheriya ezithile yaye uyavuma ukuba kubopheleleka imithetho ethile (umzekelo, ukongamela evela IRB kunye nezicwangciso zokukhusela data) . Le ndlela esibiyelweyo umyezo ibonelela ezininzi iinzuzo ukukhululwa bakulibale na oluncinci. Kakade ke, indlela esibiyelweyo umyezo idala imibuzo-abaninzi kufuneka ukufikelela, phantsi kweziphi iimeko, ixesha elide kangakanani, ngubani alihlawule ukugcina lilinde nesitiya esibiyelweyo njalo-kodwa ayithethi ukuba ayinakuphunyezwa. Enyanisweni, kukho sele ukusebenza iigadi ezibiyelweyo endaweni yokuba abaphandi Ungasebenzisa ngoku, ezifana yokugcina lwedatha Inter-University Consortium for Research Political nezeNtlalo kwiYunivesithi yaseMichigan.

Ngoko, apho kufuneka idata evela kuphononongo yakho kumanqwanqwa okuqhubeka akukho yokwabelana, igadi onodonga, ndimkhulule bakulibale? Oku kuxhomekeka iinkcukacha data yakho; nabaphandi kufuneka ukulungelelanisa Intlonelo Abantu, Beneficence, Justice, Nokuhlonela Law and Public Inzala. Xa kuhlolwa imali efanelekileyo ezinye izigqibo abaphengululi bafuna icebiso nokuvunywa IRBs, kwaye khupha data kunokuba nje enye babeyinxalenye yaloo nkqubo. Ngamanye amazwi, nangona abanye abantu bacinga woyekelelo data njenge morass ezisesikweni ndingenathemba, sele kufuneka iinkqubo ezikhoyo ukunceda abaphandi ukulungelelanisa ezi hlobo yokuziphatha ekhokelela engxakini.

Enye indlela lokugqibela ucinga ngokuchazela abanye idatha nekanga. Rhoqo ngonyaka iimoto banoxanduva amawaka abantu, kodwa musa ukuzama ukuvala ukuqhuba. Enyanisweni, umnxeba enjalo umlomo ukuqhuba ngqiqweni ngenxa yokuba ukuqhuba lwenza izinto ezininzi ezintle. Kunoko, uluntu ubeka imiqathango ngubani ukuqhuba (umzekelo, kufuneka ube neminyaka ethile, kufuneka ukuba kudlule iimvavanyo ezithile) nangendlela ukuqhuba (umzekelo, ngaphantsi kwesantya). Society Kwakhona abantu uxanduva ukunyanzelisa le mithetho (umzekelo, amapolisa), kwaye kubohlwaya abantu abo babanjiweyo nokungayithobeli. Olu hlobo lunye yokucinga elungeleleneyo ukuba uluntu isebenza olawula ukuqhuba kwakhona emalifakwe yokwabelana data. Oko kukuthi, kunokuba ukwenza iingxoxo absolutist okanye ngokuchasene yokwabelana data, ndicinga iingenelo inkulu iya kuvela kungona befunda indlela ukwabelana data ngakumbi ngokukhuselekileyo.

Ukugqibezela, umngcipheko ngolwazi yande kakhulu, yaye kunzima kakhulu ukuqikelela kunye nobuninzi bazo. Ngoko ke, kuya kuba kuhle ukuba sicinge ukuba yonke idatha ezinokuba achongeka kwaye ezinokuba novakalelo. Ukunciphisa umngcipheko yolwazi ngoxa ukwenza uphando, abaphandi olunokudala ukulandela isicwangciso ukhuseleko data. Ngaphezu koko, umngcipheko ngolwazi akuthinteli abaphandi ukusuka ngokwabelana data nezinye izazinzulu.