6.6.2 Ukuqonda nokulawula umngcipheko zolwazi

Umngcipheko wolwazi ngumngcipheko oqhelekileyo uphando loluntu; liye landa kakhulu; kwaye kuyingozi kakhulu ukuyiqonda.

Umngeni wesibini wokuziphatha ngokubhekiselele kuphando lwe-digital-age ingozi yengozi , inokulimaza ekuveliseni ulwazi (National Research Council 2014) . Ulwaphulo olunolwazi olusuka ekudaleni ulwazi lwentlalo luya kuba qoqosho (umzekelo, ukulahlekelwa ngumsebenzi), intlalo (umzekelo, ukuhlazeka), ngokwengqondo (umz., Ukudakumba), okanye ulwaphulo-mthetho (umzekelo, ukubanjwa ngenxa yokuziphatha okungekho mthethweni). Ngelishwa, ubudala bedijithali bunyusela umngcipheko wolwazi ngokubanzi-kukho ulwazi olungakumbi malunga nokuziphatha kwethu. Kwaye kunobungozi obuninzi bokungqinela kubonakalisile kunzima ukuqonda nokulawula xa kuthelekiswa nemingcipheko eyayinenkxalabo kuphando loluntu lwe-analog, njengengozi yomzimba.

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 wokungaphumeleli "kweengxelo" zivela ekupheleni kweminyaka yama-1990 eMassachusetts (Sweeney 2002) . Ikomishoni ye-Intshurensi yeqela (i-GIC) yinkampani karhulumente ejongene nokuthenga i-inshorensi yempilo kubo bonke abasebenzi baseburhulumenteni. Ngalolu msebenzi, i-GIC iqokelele iirekhodi zempilo ngokubanzi malunga namawaka abasebenzi baseburhulumenteni. Kwimizamo yokukhuthaza uphando, i-GIC yanquma ukukhulula ezi ngxelo kubaphandi. Nangona kunjalo, abazange bahlanganyele yonke idatha yabo; kunoko, "babonisa" le data ngokususa ulwazi olunjengamagama kunye needesi. Nangona kunjalo, bashiya ezinye iinkcukacha abacinga ukuba zinokubaluleka kubaphandi njengolwazi lomntu (i-zip code, usuku lokuzalwa, ubuhlanga, kunye nesondo) kunye nolwazi lwezokwelapha (ukuvakashela idatha, ukuxilongwa, inkqubo) (umzekeliso 6.4) (Ohm 2010) . Ngelishwa, le "ntengiso" ayinelanga ukukhusela idatha.

Umzekeliso 6.4: "Inkcazo" yinkqubo yokususa ulwazi olucacileyo. Ngokomzekelo, xa ukhulula iirejista zonyango zonyango zabasebenzi baseburhulumenteni, iKhomishoni ye-inshurensi yaseMassachusetts (GIC) isuse amagama kunye needesi kwiifayile. Ndisebenzisa amanqaku okucaphuna ngegama elithi "ukunyanzelisa" ngenxa yokuba inkqubo inika ukubonakala komntu ongaziwayo kodwa engaziwayo.

Ukubonisa iimpazamo ze-GIC "ukunyanzelisa", uLatanya Sweeney-ke umfundi ophumelele kwi-MIT-ehlawulwe i-$ 20 ukufumana iirekhodi zokuvota ezivela kwisixeko saseCambridge, idolophu yaseMassachusetts uWilliam Weld. Ezi rekhodi zavota zibandakanya ulwazi olunjengegama, idilesi, ikhowudi ye-zip, usuku lokuzalwa, kunye nesini. Inyaniso yokuba ifayili yedatha yezokwelapha kunye nefayile yevoti ekwabelwana ngayo i-zip code-zip, usuku lokuzalwa, kunye nesondo-kwakuthetha ukuba u-Sweeney angabadibanisa. U-Sweeney wayesazi ukuba usuku lokuzalwa lukaWeld lwangoJulayi 31, 1945, kwaye iirekhodi zokuvota ziquka abantu abathandathu kuphela eCambridge kunye naloo suku lokuzalwa. Ukongezelela, kwabo bantu abathathu, kuphela abathathu babesilisa. Yaye, kulawo madoda amathathu, kuphela enye eyabelana ngekhodi ye-Weld ye-zip. Ngaloo ndlela, idatha yokuvota ibonise ukuba nabani na kwiinkcukacha zonyango kunye nentsebenziswano yeWeld yokuzalwa, isini kunye ne-zip code nguWilliam Weld. Ngokwenene, ezi zintathu zengcaciso zinikezela ngeminwe ecacileyo kuye kwidatha. Ukusebenzisa le nyaniso, u-Sweeney wakwazi ukufumana iirekhodi zezokwelapha zeWeld, kwaye, ukuba amxelele ngaye, wamthumelela ikopi yeengxelo zakhe (Ohm 2010) .

Umzobo 6.5: Ukuzibandakanya kwakhona kwedatha "engabonakaliyo". ULatanya Sweeney wadibanisa iirekhodi zempilo kunye neengxelo zokuvota ukwenzela ukuba afumane iirekhodi zezogulana zikaRubhulane uWilliam Weld Ulungelelaniso oluvela ku- Sweeney (2002) , isiqendu 1.

Umsebenzi ka-Sweeney ubonisa isiseko esisisiseko sokuhlaselwa ngokutsha-ukuba athathe ixesha elivela kuluntu loluntu lwezokuphepha. Kule hlaselo, iiseti ezimbini zeedatha, akukho nto ngokwayo ezityhila ulwazi olubucayi, zidibaniswa, kwaye ngolu hlobo, ulwazi olubukhali luyabonakala.

Ekuphenduleni umsebenzi we-Sweeney, kunye nomnye umsebenzi odibeneyo, abaphandi ngoku ngokubanzi basusa ulwazi oluninzi-konke okubizwa ngokuba "ulwazi oluchongiweyo" (PII) (Narayanan and Shmatikov 2010) -ukunyusa inkqubo "yobungqina." Ngaphezu koko, abaphandi abaninzi Ngoku uyaqonda ukuba iinkcukacha ezithile-njengamarekhodi ezonyango, iirekhodi zezemali, iimpendulo kwimibandela yokuhlola malunga nokuziphatha okungekho mthethweni-mhlawumbi iyakhathazeka ukukhulula nangemva kokuba "imbonakalo." Nangona kunjalo, imizekelo endikuyo kubonisa ukuba abaphandi bezentlalo kufuneka ukutshintsha ukucinga kwabo. Njengenyathelo lokuqala, kuya kuba bubulumko ukucinga ukuba zonke iinkcukacha zisenokuba abonakalayo kwaye yonke idatha ekungenzeka ibe nobuzaza. Ngamanye amagama, kunokuba sicinge ukuba ingozi yolwazi isebenza kwicandelo elincinane leeprojekthi, kufuneka sicinge ukuba isebenza-kwinqanaba elithile-kuzo zonke iiprojekthi.

Zomibini iinkalo zolu hlaziyo ziboniswa ngumvuzo weNetflix. Njengoko kuchazwe kwisahluko 5, iNetflix ikhishwe izilinganiso zee-movie ezigidi eziyi-100 ezinikezwa ngamalungu angaba ngu-500,000, kwaye inefowuni evulekileyo apho abantu bevela kulo lonke ihlabathi bangenise i-algorithms enokuphucula ikhono likaNetflix lokuncoma ama-movie. Ngaphambi kokuba ukhulule idatha, i-Netflix isusile nayiphi na inkcazelo ebonakalayo ngokubonakalayo, njengamagama. Baye bathatha inyathelo elongezelelweyo baze bavelise amathintelo ambalwa kwezinye iirekhodi (umzekelo, ukutshintsha ezinye izilinganiso ukusuka kwiinkwenkwezi ezine ukuya kwiinkwenkwezi ezi-3). Kodwa kungekudala, bafumanisa ukuba nangona benzame, idatha ayengasichazwanga.

Kwiiveki ezimbini emva kokukhishwa kwedatha, u-Arvind Narayanan noVitaly Shmatikov (2008) babonisa ukuba kwakunokwenzeka ukuba bafunde ngabantu abakhethiweyo be-movie. Ubuqhetsepho ekuhlaseleni kwabo ukuchazwa kwakhona kwafana no-Sweeney's: ukudibanisa kunye nemithombo yolwazi emibini, enye eneenkcukacha ezinobungozi kunye nolwazi olungabonakaliyo nolwazi oluqulethwe ngabantu. Inye yale mithombo yolwazi ingaba yindawo ephephile ngabanye, kodwa xa idibeneyo, i-dataset edibeneyo inokudala umngcipheko wolwazi. Kwimeko ye-Netflix yedatha, nantsi indlela enokwenzeka ngayo. Khawucinge ukuba ndikhetha ukwabelana ngeengcamango zam ngezenzo kunye nokudlala amafilimu kunye nabasebenzi bam basebenzisana nabo, kodwa ukuba ndikhetha ukungaboni ngolwazi lwam malunga namafilimu enkolo nezopolitiko. Abasebenzi bam basebenzisane nabo bangasebenzisa ulwazi endabelane nabo ukufumana iirekhodi zam kwi-Netflix idatha; ulwazi ondibelana ngalo lunokuba yiminwe yangasese efana nomhla wokuzalwa kukaWilliam Weld, ikhowudi ye-zip kunye nesondo. Emva koko, ukuba bafumene impepha yam fingerprint yangasese kwidatha, banokufunda ukulinganisa kwam zonke iifrimu, kubandakanywa namafilimu endikhethayo ukuba ndingabelane ngawo. Ukongezelela kulolu hlobo lokuhlaselwa okujoliswe kumntu oyedwa, uNarayanan noShmatati bakubonisa ukuba kunokwenzeka ukwenza ukuhlaselwa ngokubanzi -kubandakanya abantu abaninzi-ngokudibanisa idatha yeNetflix kunye nedatha yokulinganiswa komntu kunye neye-movie abaye bayinyula ukuthumela kwi-Movie Database ye-Intanethi (IMDb). Ngokulula, naluphi na ulwazi oluyimpinci eyodwa yeminwe kumntu othize-nokuba isethi yabo yokulinganiswa kwe-movie-ingasetyenziselwa ukuchonga.

Nangona idatha ye-Netflix ingakwazi ukuphinda iphinde ibonelelwe kuyo nayiphi na into ejoliswe kuyo okanye ibanzi, iyakubonakala ibonakala ibe yingozi. Emva kwakho konke, ukulinganiswa kwe-movie akubonakali kunzima kakhulu. Nangona oko kunokuba yinyaniso ngokubanzi, kwabanye abantu abangama-500,000 kwi-dataset, ukulinganiswa kwe-movie kungabonakali. Enyanisweni, ekuphenduleni ukubonakaliswa kwakhona, umfazi onqamlekileyo wesibini ujoyine ist-action action against Netflix. Nantsi indlela inkcazo yaboniswa ngayo kumgwebo wabo (Singel 2009) :

"[M] ovie kunye nedatha yokulinganisa iqulethe ulwazi lwe ... uhlobo oluthile noluntu olubucayi. Idatha ye-movie yelungu ibonisa umdla womntu siqu kaNetflix kunye / okanye ulwa nemibandela eyahlukeneyo yabantu, kubandakanya ukuziphatha ngokwesondo, ukugula kwengqondo, ukuphulukana nokusela utywala, nokuxhatshazwa kwintsholongwane, ukuphathwa kakubi ngokomzimba, ulwaphulo lwasekhaya, ukukrexeza nokudlwengulwa. "

Ukuchongwa kwakhona kweDatha yeNewflix Idatha kubonisa zombini ukuba yonke idatha ingabonakala kwaye yonke idatha ingenakukhathazeka. Kule ngongoma, unokucinga ukuba oku kusebenza kuphela kwiedatha ezicinga ukuba ngabantu. Okumangalisa kukuba, akunjalo. Ekuphenduleni isicelo soMthetho weNkululeko yeNgcaciso, uRhulumente weSixeko saseNew York wakhupha iirekhodi zonke zokuhamba ngetekisi eNew York ngowama-2013, kubandakanywa iipopu kunye nokulahla amaxesha, indawo kunye neendleko zokuhamba (khumbula kwisahluko 2 ukuba u- Farber (2015) sebenzisa idatha efanayo ukuvavanya iingcamango ezibalulekileyo kwezoqoqosho lwezabasebenzi). Ezi nkcukacha malunga nohambo lweeteksi zingabonakala zingenangqondo kuba zingabonakali ukubonelela ngolwazi ngabantu, kodwa u-Anthony Tockar waqonda ukuba le dataset yenene iqulethe ulwazi oluninzi olukhutheleyo ngabantu. Ngokomzekelo, wayejonge zonke iindwendwe eziqala kwi-Hustler Club-iqela elikhulu le-strip eNew York-phakathi kobusuku bobusuku kunye no-6 ekuseni kwaye bafumana indawo yabo yokukhupha. Olu phando luchazwe-ngoqobo-uluhlu lweedesi zabanye abantu abaye bahlala kwiHustler Club (Tockar 2014) . Kunzima ukucinga ukuba urhulumente weso sixeko ubenokukukhumbuza xa ukhululile idatha. Enyanisweni, le nkqubo ingasetyenziselwa ukufumana iidilesi zasekhaya zabantu abavakatye nayiphi na indawo kwisixeko-ikliniki yezokwelapha, isakhiwo sikaRhulumente okanye isikhungo senkolo.

Ezi zimbini Iingxelo zeNetflix kunye neNew York City yeeteksi zedatha zibonisa ukuba abantu abanezakhono abanokukwazi ukulinganisa ngokufanelekileyo ingozi yolwazi kwi-data abayikhululayo-kwaye le meko ayinalo ikhethekileyo (Barbaro and Zeller 2006; Zimmer 2010; Narayanan, Huey, and Felten 2016) . Ukongezelela, kwiimeko ezininzi ezinjalo, iinkcukacha ezinengxaki zifumaneka ngokukhululekileyo kwi-intanethi, ezibonisa ubunzima bokususa ukukhutshwa kwedatha. Ngokuqokelela, le mizekelo-kunye nophando kwi-computer yesayensi malunga nobumfihlo-kubangele kwisigqibo esibalulekileyo. Abaphandi kufuneka sicinge ukuba zonke iinkcukacha zisenokuba abonakalayo kwaye yonke idatha ekungenzeka ibe nobuzaza.

Ngelishwa, akukho isisombululo esilula kwiinyaniso ukuba zonke iinkcukacha ziyakwazi ukuchonga kwaye zonke iinkcukacha zinokuthi zivelele. Nangona kunjalo, enye indlela yokunciphisa ingozi yolwazi ngexesha usebenza kunye nedatha kukudala nokulandela isicwangciso sokukhuselwa kwedatha . Esi sicwangciso siya kunciphisa ithuba lokuba idatha yakho iya kuvuza kwaye iya kunciphisa ingozi xa ukuvuza kuvela ngandlela-thile. Izicwangciso zokukhuselwa kwedatha, njengoluhlobo lwe-encryption oluyisebenzisayo, luya kutshintshwa ngokuhamba kwexesha, kodwa i-UK Data Services inceda ngokucwangcisa izicwangciso zesicwangciso sokukhuselwa kwedatha kwiintlobo ezintlanu ezizibiza ngokuba zikhuseleko ezintlanu : iiprojekthi ezikhuselekileyo, abantu abakhuselekileyo , izicwangciso ezikhuselekileyo, idatha ekhuselekileyo, kunye neziphumo ezikhuselekileyo (itafile 6.2) (Desai, Ritchie, and Welpton 2016) . Akukho nanye kwiindawo ezikhuselekileyo zintlanu zibonelela ngokukhuselekileyo. Kodwa kunye benza iimeko ezinamandla ezinokunciphisa ingozi yolwazi.

Ukongezelela ekukhuseleni idatha yakho ngenkathi usebenzisa, isinyathelo esinye kwinkqubo yophando apho ingozi yolwazi ingakumbi ngokubelana ngolwazi nabanye abaphandi. Ukwabelana ngolwazi phakathi kwezenzululwazi yintengo ebalulekileyo yenzululwazi, kwaye iququzelela kakhulu ukuqhubela phambili kolwazi. Nantsi indlela iNdlu yaseMelika yaseMelika ichaza ukubaluleka kokwabelana ngolwazi (Molloy 2011) :

"Ukufikelela kwedatha kuyimfuneko ukuba abaphandi bavelise, baqinisekise kwaye bakhiwe kwiziphumo ezichazwe kwincwadi. Ukuxhomekeka kufuneka ukuba kukuba, ngaphandle kokuba kukho isizathu esinamandla, idatha kufuneka ibhengezwe ngokupheleleyo kwaye yenziwe ngokubanzi. "

Nangona kunjalo, ngokukwabelana ngolwazi lwakho kunye nomnye umphandi, unokwandisa umngcipheko wolwazi kubathathi-nxaxheba bakho. Ngaloo ndlela, kubonakala ngathi ukwabelana kwedatha kubangela ukungqinelana okubalulekileyo phakathi koxanduva lokukwabelana ngolwazi kunye nezinye izazinzulu kunye noxanduva lokunciphisa umngcipheko wolwazi kubathathi-nxaxheba. Ngethamsanqa, le ngxaki ayinzima kakhulu njengoko ibonakalayo. Kunoko, kungcono ukucinga malunga nokwabelana ngeedatha njengoko kuwela kwi-continuum, kunye neengongoma nganye kulolu hlobo oluqhubekayo lunikeza umxube ohlukeneyo lweenzuzo kuluntu kunye nobungozi kubathathi-nxaxheba (umzekeliso 6.6).

Ngenye indlela, ungabelana ngolwazi lwakho ngaphandle komntu, olunciphisa ingozi kubathathi-nxaxheba kodwa luyancipha inzuzo kuluntu. Kwesinye isiqhelo, unako ukukhupha kwaye uyilibale , apho idatha "ibonakaliswa" kwaye ithunyelwe kuwo wonke umntu. Ngokumalunga nokukhulula idatha, ukukhulula nokulibala ukunikezelwa kwezibonelelo eziphezulu kuluntu kunye nomngcipheko ophezulu kubathathi-nxaxheba. Phakathi kwezi zihlandlo zibini ezigqithiseleyo ziluhlu lwee-hybrids, kubandakanywa noko ndiza kubiza indlela yokulima engadini . Ngaphantsi kwendlela, idatha ikwabelwana nabantu abahlangabezana nemigangatho ethile kwaye bavuma ukuboshwa ngemithetho ethile (umz., Ukubhekwa kwe-IRB kunye nesicwangciso sokukhuselwa kwedatha). Indlela yokudibanisa idibaniso inikeza ezininzi iinzuzo zokukhululwa kwaye zikhohlwe ngengozi encinane. Ewe, indlela enjalo idala imibuzo eninzi-ngubani na onokufikelela, phantsi kweyiphi imimiselo, kunye nexesha elide, ngubani omele ahlawule ukuze agcine kwaye apolise umyezo odibeneyo, njl. -Kodwa oku akunakwenzeka. Enyanisweni, kukho iindidi ezisemagqabini kwindawo esetyenziswa ngabaphandi ngoku, njengengxelo yolwazi ye-Inter-yunivesithi Consortium yoPhando lwezoPolitiko nezeNtlalo kwiYunivesithi yaseMichigan.

Umzobo 6.6: Amasu okukhululwa kwedatha awela phantsi kokuqhubeka. Lapho ufanele ube khona kulolu qhubela phambili kuxhomekeka kwiinkcukacha ezithile zeenkcukacha zakho, kwaye ukuphononongwa kwenkxaso yesithathu kunokukunceda ukuba wenze isigqibo malunga nomngcipheko ofanelekileyo kunye nenzuzo kwimeko yakho. Ubume obufanelekileyo beli khoba lixhomekeke kwiinkcukacha zedatha kunye nophando (Goroff 2015) .

Ngoko, iphi idatha ekufundwa kwakho ekufuneka iqhubekele ekuqhubekeni kokungahlanganyeli, ingadi, kunye nokukhululwa kwaye uyibale? Oku kuxhomekeke kwiinkcukacha zedatha yakho: abaphandi kufuneka balinganise ukuhlonipha abantu, ukuxhamla, ubulungisa, kunye nokuhlonipha uMthetho noMdla woluntu. Kubonwa ngolu hlobo, ukwabelana kwedatha akuyona into eyahlukileyo yokuziphatha; yinto enye yezinto ezininzi zophando apho abaphandi kufuneka bafumane ibhalansi efanelekileyo yokuziphatha.

Abanye abagxeki ngokubanzi bachasene nokwabelana kwedatha kuba, ngokombono wam, bagxile kwimingcipheko-okuqinisekileyo ukuba bayinyani-kwaye bayayihoxisa izibonelelo zayo. Ngoko, ukuze sikhuthaze kugxininiso kwimingcipheko kunye neenzuzo, ndingathanda ukunikela ngomfanekiso. Unyaka ngamnye, iimoto zinoxanduva lokufa kwabangamawaka, kodwa asizami ukuvala ukuqhuba. Enyanisweni, umnxeba wokuvinjelwa ukuqhubela ukuqhubela ukuqhubela umbane kungeyinto engekho nto kuba ukuqhuba kudala izinto ezininzi ezintle. Kunoko, uluntu lubeka izithintelo kulowo onokuqhuba (umzekelo, imfuno yokuba ngumdala othile kwaye uphumelele iimvavanyo ezithile) nendlela anokuqhuba ngayo (umzekelo, phantsi komda wokukhawuleza). Umbutho unabantu abanomsebenzi wokunyanzelisa le mithetho (umz., Amapolisa), kwaye siyabahlwaya abantu ababanjwayo. Uhlobo olufanayo lokucinga ngokulinganiselayo uluntu olusebenzayo ekulawuleni umqhubi lunokusetyenziswa kwi-data share. Oko kukuthi, kunokuba senze iingxabano ze-absolutist okanye malunga nokwabelana ngedatha, ndicinga ukuba siya kuqhuba inkqubela phambili ngokugxila kwindlela esinokunciphisa ngayo ingozi kunye nokwandisa inzuzo kwi-data ukwabelana.

Ukugqiba, umngcipheko wolwazi uye wanda kakhulu, kwaye kunzima ukuqikelela nokulinganisa. Ngoko ke, kukulungele ukucinga ukuba zonke iinkcukacha ziyakwazi ukuchonga kwaye ziyakwazi ukukhathazeka. Ukunciphisa umngcipheko wolwazi ngexesha lophando, abaphandi bangadala kwaye balandele isicwangciso sokukhuselwa kwedatha. Ukongeza, ingozi yolwazi ayithinteli abaphandi ukuba babelane ngolwazi nabanye ososayensi.