- Ubuhlakani bokwenziwa bukwenza kube nokwenzeka ukubona, ukuhlanganisa nokuphendula ezinsongweni ezinkulu ze-cyber, kunciphisa imiphumela engamanga kanye nezikhathi zokusabela.
- I-AI ekhiqizayo ithuthukisa kokubili ukuzivikela (ukulingisa, idatha yokwenziwa, ukuzenzekela) kanye nokuhlasela (ubugebengu bokweba imininingwane ebucayi obuthuthukisiwe, ama-deepfake, ukwenziwa kwezwi).
- Ukufunda komshini kusetshenziswa ekuhlukanisweni kwedatha, ekuhlaziyweni kokuziphatha, ekufakweni kwephrofayela komsebenzisi, nasekuvinjweni kwe-bot, okuthuthukisa ukuvikelwa ngaphandle kokufaka esikhundleni amaqembu abantu.
- Impumelelo yesikhathi esizayo incike ekuqinisekiseni ipayipi le-AI uqobo, ukuhambisana nemithethonqubo yedatha, kanye nokuhlanganisa ukuzenzekela kanye nokuqapha kanye nokwahlulela kwabantu.
Ezweni elisebenzisa kakhulu idijithali, ukuphepha kwe-inthanethi sekuyinto ebalulekile yokuphepha futhi kuyisihluthulelo sokuphepha nobumfihlo enkathini yedijithali yamabhizinisi, ohulumeni, kanye nezakhamuzi ezivamile. Yonke insizakalo entsha yamafu, yonke idivayisi exhunyiwe, kanye nohlelo lokusebenza ngalunye esilufakayo lwandisa indawo yokuhlasela izigebengu ze-inthanethi ezingayisebenzisa.
Ngesikhathi esifanayo, ukufika kobuhlakani bokwenziwa (AI), ukufunda komshini (ML), kanye ne-AI yokukhiqiza kushintshe ngokuphelele umdlalo. Lobu buchwepheshe abugcini nje ngokuqinisa ukuzivikela kodwa futhi busetshenziswa abahlaseli ukuqalisa imikhankaso emikhulu, enembile, nenzima ukuyibona, okwenza kube kubalulekile ukuqonda amakhono abo, indlela abasebenza ngayo, nokuthi imikhawulo yabo ikuphi.
Indlela i-AI eguqula ngayo ukuphepha kwe-cyber
I-AI ilethe intuthuko ephawulekayo endleleni izigameko ezitholakala ngayo, eziphenywa ngayo, futhi eziphendulwa ngayo , ikakhulukazi ezindaweni lapho izigidi zezehlakalo zokuphepha zenziwa khona nsuku zonke. Amapulatifomu afana ne-SIEM, i-XDR, i-NDR, noma izixazululo zesimanje ze-endpoint bezingeke ziphathwe ngaphandle kwama-algorithms akwazi ukuhlunga umsindo nokubeka phambili lokho okubaluleke kakhulu.
Ezinhlanganweni eziningi, izinhlelo zokuphepha zibika izinkulungwane zezenzakalo njalo ngeminithi : ukuxhumana okungajwayelekile, ukungena ngemvume okuphindaphindiwe, ukulanda okusolisayo, izinguquko zokucushwa, njalo njalo. Iningi lalezi zixwayiso alinangozi, kodwa ezimbalwa zifihla ukuziphatha okunonya ngokusobala. Yilapho i-AI ikhanya khona, njengoba ifunda ukuhlukanisa amaphethini asemthethweni kulawo akhomba ukuhlaselwa kwangempela.
Amamodeli okufunda komshini ahlobanisa imisebenzi ebonakala ingenangozi, uma ibukwa ngokwehlukana, (ukungena ngemvume ngemva kwamahora okusebenza, ifayela elicindezelwe, ukufinyelela kuseva ethile), kodwa okuhlangene kwakha umzila ojwayelekile we-ransomware, ukunyakaza okuseceleni, noma ukukhishwa kwedatha; yingakho ukuba nama-backups endawo kubalulekile.
Ngaphezu kwalokho, izixazululo ezithuthuke kakhulu zihlanganisa izinjini ze-AI ezikhiqizayo ezikwazi ukubhala imibiko eqondakalayo ngolimi lwemvelo , zifingqa okwenzekile, umthelela ongaba khona, ukuthi yiziphi izinhlelo ezithintekile, nokuthi yiziphi izenzo ezinconywayo. Lokhu kunciphisa kakhulu isikhathi sokuhlaziya futhi kwenza kube lula kubaphathi abangebona ochwepheshe ukuqonda ubungozi nokwenza izinqumo.
Omunye umnikelo obalulekile ukuhlonza ngokuzenzakalelayo ubuthakathaka kanye nezimpahla ezingaziwa : amadivayisi axhuma kunethiwekhi ngaphandle kwemvume, izinhlelo zokusebenza zamafu ezingafakwanga ohlwini, izinhlelo zokusebenza ezingakalungiswa, noma idatha ebucayi engavikelekile kahle. Ngokuqhathanisa izinqolobane, ukuhamba kwenethiwekhi, kanye nezinqubomgomo, i-AI iveza izikhala zokuphepha ezazingatholakali ngaphambilini.
Futhi isibe umngane oqondile wamaqembu e-SOC, njengoba ihumusha imibuzo eyinkimbinkimbi kanye nemiphumela yobuchwepheshe ibe ulimi oluvamile . Abahlaziyi abancane bangaphenya izehlakalo ngaphandle kokufunda izilimi zemibuzo ezithuthukisiwe, futhi ithuluzi ngokwalo lisikisela izinyathelo zokulungisa, iziqondiso zokulawula ukuhlaselwa, kanye nemikhuba emihle yokuvimbela ukuphinda kwenzeke.
Ngokuhlanganisa nokuhlaziya idatha evela emithonjeni eyahlukahlukene—amalogi okuphepha, ithrafikhi yenethiwekhi, ubuhlakani bezinsongo zangaphandle, ukuziphatha komsebenzisi, kanye nama-endpoints— i-AI inikeza umbono ohlangene wesimo sokuphepha , okuhlanganisa nokuphathwa kwemishini yenethiwekhi , iqokomisa amaphethini okuhlasela okungenakwenzeka ukuwabona ngesandla. Leli khono lokuhlanganisa liguqula idatha ephithizelayo ibe ulwazi olusebenzayo ngempela.
Indawo eyodwa lapho i-AI yenza khona umehluko omkhulu ukunciphisa okuhle okungamanga kanye nokubi okungamanga . Ngokubona amaphethini, ukuhlaziywa komongo, ukutholwa okungavamile, kanye nokufunda okuqhubekayo, amamodeli alungisa ukuzwela kwawo ukuze anciphise kokubili izexwayiso ezingafanele kanye nezinsongo ezingabanjwanga—okubalulekile ekulweni nokukhathala kokuxwayisa phakathi kwabasebenzi bezokuphepha.
Okokugcina, i-AI inikeza ukusabalala okungenakufaniswa nomsebenzi womuntu kuphela . Ingacubungula ukusakazwa kwedatha okukhulu ngesikhathi sangempela, ifunde esigamekweni ngasinye, futhi izivumelanise namaqhinga amasha okuhlasela. Njengoba inani lezinsongo ze-cyber kanye nobunzima bengqalasizinda kukhula, leli khono lokukhula ngaphandle kwezindleko zabasebenzi ezikhuphukayo liba yinto ebaluleke kakhulu.
Ukusetshenziswa okusebenzayo kwe-AI ekuphepheni kwe-cyber
Empeleni, i-AI isivele ikhona cishe kuzo zonke izingqimba zokuzivikela kwenhlangano. Kusukela ekuqinisekisweni komsebenzisi kuya ekutholeni ukuziphatha okungavamile, indima yayo idlulela ngale kokuba "okungeziwe" kobuchwepheshe obulula.
Ekuphathweni kobunikazi, isibonelo, i-AI isiza ukuqinisa ukuvikelwa kwephasiwedi kanye nokuqinisekiswa ngokuthola ukusetshenziswa okungajwayelekile, ukufinyelela ezindaweni ezingavamile, noma amadivayisi angakaze abonwe ngaphambili, ikakhulukazi ezindaweni zeselula njengokuphepha kwe-Android vs. iOS . Iphinde ifake isandla ezinhlelweni zokuqinisekiswa eziguquguqukayo, iphakamisa izinga lokuphepha lapho okuthile "kungahlangani" kuphethini yomsebenzisi.
Emkhakheni wokukhwabanisa nokutholwa nokuvimbela ukwebiwa kobunikazi (ubugebengu bokweba imininingwane ebucayi, ubugebengu bokweba imininingwane eyimfihlo, ubugebengu bokweba imininingwane eyimfihlo, ubugebengu bokweba imininingwane eyimfihlo, ubugebengu bokweba imininingwane eyimfihlo, ubugebengu bokweba imininingwane eyimfihlo, ubugebengu bokweba imininingwane eyimfihlo, ubugebengu bokweba imininingwane eyimfihlo, ubugebengu bokweba imininingwane eyimfihlo, ubugebengu bokweba imininingwane eyimfihlo, ubugebengu bokweba imininingwane eyimfihlo, kanye nobugebengu bokweba imininingwane eyimfihlo ukuze kuhlukaniswe ukuxhumana okusemthethweni nemizamo eyinkimbinkimbi yokukhohlisa ngenxa ye-AI ekhiqizayo, futhi kuyingxenye ebalulekile yokuvikelwa kwe-inthanethi.
Ukuphathwa kobungozi kanye nokuphepha kwenethiwekhi nakho kuyazuzisa kakhulu. Izinjini zokufunda ngomshini zibeka phambili amaphutha okuphepha ngokusekelwe ekusebenziseni kwazo kwangempela kanye nomongo othize wenhlangano, kuyilapho izinhlelo ezisekelwe ku-AI ziqapha ithrafikhi ukuthola amaphethini angajwayelekile, ukuxhumana nezizinda ezinonya, noma ukunyakaza okuseceleni phakathi kwamaseva, futhi ziphatha okhiye ngamamojula okuphepha kwehadiwe.
Ukuhlaziywa kokuziphatha sekuyinto ebalulekile: amaphrofayili okuziphatha akhelwe abasebenzisi nezinhlelo , ukuze noma yikuphi ukuphambuka okufanele—izikhathi ezingajwayelekile, ukufinyelela okungajwayelekile kudatha ebucayi, amavolumu okulanda angajwayelekile—kuqalise isexwayiso noma ngisho nempendulo ezenzakalelayo.
Amathuluzi okuphepha kwe-inthanethi asebenza nge-AI
Lo mbono muhle futhi muhle, kodwa umthelela wangempela ubonakala ezixazululweni eziqondile esezivele zihlanganisa i-AI noma i-ML njengengxenye eyinhloko yokusebenza kwazo. Phakathi kwezibaluleke kakhulu, singaqokomisa amaqembu amaningana kanye neminye imikhiqizo emele isigaba ngasinye.
Okokuqala, sinezixazululo zokuphepha ze-endpoint ezisebenzisa i-AI , ezikwazi ukuvimba i-malware engaziwa ngokuhlaziya ukuziphatha kwayo ngesikhathi sangempela, ngaphandle kokuthembela kuphela ezimpawini. Ama-antivirus amaningi esizukulwane esilandelayo afaka lezi zinjini, ahlanganisa ukuhlaziywa okungaguquki nokuguquguqukayo namamodeli okubikezela.
Ama-firewall esizukulwane esilandelayo (ama-NGFW) asebenzisa i-AI ahlinzeka ngokuhlolwa okujulile kwethrafikhi, ukuhlonza uhlelo lokusebenza, ukutholwa kokungena, kanye nokuhlukaniswa okuhlakaniphile. I-AI isiza ekutholeni amaphethini okuxhumana angavamile, imigudu eyimfihlo, noma imizamo yokugwema inqubomgomo engaphuthelwa yi-firewall yendabuko. Ngezakhiwo zomphetho kanye nokwahlukaniswa , kunconywa ukuhlaziywa kwe-router.
Ngaphakathi kokuqapha okuhlanganisiwe, amapulatifomu e -SIEM (Ulwazi Lokuphepha kanye Nokuphathwa Kwemicimbi) aguqukele ezinjinini zokuhlaziya ezihlakaniphe kakhulu. Ahlobanisa imicimbi evela emithonjeni eminingi, asebenzise amamodeli okuziphatha, futhi abeke phambili izehlakalo ezisolisayo, anciphise umthwalo womsebenzi wezandla wama-SOC.
Izixazululo zokuphepha kwamafu ezisebenzisa i-AI nazo zithole ukuthandwa , ziqapha izindawo ze-IaaS, i-PaaS, kanye ne-SaaS, zithola ukulungiselelwa okungalungile, ukufinyelela okungavamile kwe-API, kanye nokunyakaza okungavamile phakathi kwezifunda noma ama-akhawunti. Ezingqalasizinda zamafu amaningi, lezi zinjini zibalulekile ekugcineni ukubonakala.
Okokugcina, kukhona amathuluzi okuthola nokuphendula inethiwekhi (NDR) asebenzisa i-AI , aklanyelwe ngqo ukuthola izinsongo ze-cyber ngokuhlaziya okujulile kwethrafikhi yenethiwekhi. Ahlonza ukuhlaselwa komugqa womyalo, ukuhlunga, ukuskena kwangaphakathi, kanye nomsebenzi we-bot, futhi anikeza izimpendulo ezenzakalelayo njengokuhlukanisa amadivayisi noma ukuvimba ukuxhumana.
I-AI Ekhiqizayo: umngcele omusha wokuphepha kwe-cyber
Ukuvela kwe -AI ekhiqizayo (njengezinhlobo ze-GPT noma ama-GAN) kuvule inkundla entsha ngokuphelele emkhakheni wokuphepha kwe-cyber. Lawa mamodeli awagcini nje ngokuhlaziya idatha, kodwa futhi ayakwazi ukukhiqiza okuqukethwe: umbhalo, izithombe, umsindo, ividiyo, noma ngisho nekhodi.
Ngasohlangothini lokuzivikela, i-AI ekhiqizayo ivumela ukulingiswa kokuhlaselwa kwe-cyber okuyinkimbinkimbi ukuze kuhlolwe ukuzivikela , ukukhiqizwa kwedatha yokwenziwa yokuqeqesha izinhlelo ngaphandle kokubeka engcupheni ulwazi lwangempela, kanye nokudalwa kwezimo zokuqeqesha ezingokoqobo kakhulu zamaqembu okusabela ezigamekweni.
Ezindaweni ze-SOC kanye namapulatifomu e-SIEM, amamodeli akhiqizayo afunda ekuziphatheni okuvamile kwenethiwekhi futhi aveze ukuphambuka okucashile okungase kubonise i-malware, i-ransomware, noma ithrafikhi efihliwe, okuthuthukisa kakhulu ukutholwa kwe-anomaly uma kuqhathaniswa nemithetho engaguquki.
Ngaphezu kwalokho, lobu buchwepheshe bunegalelo ekwenziweni kwemisebenzi yokuphepha ngokuzenzakalela okuthuthukisiwe : kusukela ekuphakamiseni imithetho ye-firewall elungiselelwe kahle kuya ekukhiqizeni izikripthi zempendulo yezehlakalo, ngisho nokubhala imibiko ecacile yabaphathi kusuka kumalogi obuchwepheshe ayinkimbinkimbi. I-AI ekhiqizayo isebenza njengohlobo lomsizi okhethekile osindisa amahora amaningi omsebenzi ophindaphindwayo.
Umthelela wayo ekuqeqeshweni nawo mkhulu kakhulu, njengoba ivumela ukuvuselelwa kwezindawo zokuhlasela ezingokoqobo ezivumelana ngokuguquguqukayo nezinga lomfundi, zihlanganisa ama-vector ahlukene (i-phishing, ukunyakaza okuseceleni, ukukhushulwa kwamalungelo, ukukhishwa) ukuqeqesha kokubili amakhono obuchwepheshe kanye nokwenza izinqumo ngaphansi kwengcindezi.
Ukuhlaselwa kwe-cyber okuqhutshwa yi-AI ekhiqizayo
Ngeshwa, izigebengu ze-inthanethi ziye zashesha kakhulu ukusebenzisa i-AI ekhiqizayo ukuze zizuze . Lapho ngaphambili babedinga isikhathi, ulwazi lobuchwepheshe, kanye nezinga elithile lamakhono okuxhumana nabantu, manje sebenamathuluzi azenza umsebenzi omningi ube ngokoqobo.
Isibonelo esicacile ngabakhiqizi bemibhalo abathuthukile , abakwazi ukubhala izindaba ezingamanga, ama-imeyili e-phishing, noma imiyalezo yokuphanga ngesiSpanishi esiphelele, ngaphandle kwamaphutha esipelingi noma ukushintshana kwamagama ngendlela exakile. Lokhu kwandisa kakhulu amathuba okukhohlisa isisulu, njengoba i-imeyili "izwakala" njengokuxhumana okusemthethweni okuvela ebhange, kunethiwekhi yokuxhumana nabantu, noma ejensi kahulumeni.
Amathuluzi okudala amavidiyo kanye nama-deepfake nawo andile , okuvumela abasebenzisi ukuthi babeke ubuso kwezinye izidumbu noma bashintshe izinkulumo namagama kuziqeshana zangempela. Ngesofthiwe ekhethekile, kungenzeka ukukhiqiza amavidiyo mbumbulu ezombusazwe, abaphathi, noma amalungu omndeni akholisa kakhulu noma ubani owatholayo.
Ukwenziwa kwe-cloning yezwi sekufinyeleleka kalula ngenxa yamamodeli, ngemizuzu embalwa nje yomsindo wangempela, angalingisa cishe ngokuphelele ithoni yezwi lomuntu, indlela akhuluma ngayo, kanye nokuma kancane. Lawa mazwi ajulile avumela izingcingo lapho kuzwakala sengathi ilungu lomndeni, umphathi wenkampani, noma umphathi webhange uyakhuluma.
Elinye lamacala akhathaza kakhulu yilelo lokukhwabanisa ngezimali kusetshenziswa izwi eliqoshwe njengelungu lomndeni . Isisulu sithola ucingo oluvela kothile ozwakala njengomntwana waso, umlingani waso, noma isihlobo esiseduze, ecela ukudluliselwa ngokushesha ngenxa yesimo esiphuthumayo esicatshangelwayo. Bebhekene nokucindezeleka ngokomzwelo kanye nokwethembeka okubonakalayo kwezwi, abaningi bagcina bekhokha kakhulu kuma-akhawunti alawulwa abahlaseli.
Umthelela we-AI ekugebeni ubugebengu kanye nobunjiniyela bezenhlalo
Ubunjiniyela bezenhlalo, obuhlanganisa wonke amasu aklanyelwe ukukhohlisa abantu nokubakholisa ukuthi benze okuthile okubalimazayo , buthole umngane oyingozi ku-AI yokukhiqiza. Lokho okwakudingeka amahora ocwaningo lwesandla manje sekungenziwa ngokuzenzakalela ngezinga elikhulu.
Ngokwesiko, ukuqala umkhankaso wobugebengu bokweba imininingwane ebucayi kwakuhilela ukuphenya ngokucophelela isisulu : isikhundla sabo, ubudlelwano, izintshisekelo, abahlinzeki, njalo njalo. Lokhu kwakubiza kakhulu futhi kuthatha isikhathi, ngakho ukuhlaselwa okuyinkimbinkimbi kwakungajwayelekile. Namuhla, i-AI ingahlola izinkundla zokuxhumana, imithombo evulekile, kanye nama-imeyili adlule ukuze yakhe amaphrofayili anemininingwane eminingi ngemizuzu embalwa.
Imikhankaso iye yahlukahluka: ngaphezu kweposi evamile, sine -SMSishing (imiyalezo yombhalo kanye nemiyalezo esheshayo) , ukukhwabanisa ezinkundleni zokuxhumana, izingcingo ezinonya (i-vishing), amadrayivu e-USB "akhohliwe" ukuze alinge umsebenzisi (ukuthanda), kanye nokusetshenziswa okuvamile kwamakhodi e-QR aguquliwe (i-QRishing), aqondisa kabusha kumawebhusayithi mbumbulu noma afake i-malware.
Ngokuhamba kwesikhathi, abahlaseli baye balungisa amaqhinga abo: kusukela emilayezweni evamile kakhulu, baye bathuthela kuma -imeyili enziwe ngendlela yomuntu siqu alingisa izinqubo zangaphakathi zangempela , ukuxhumana okuvela kubaphathi noma abahlinzeki abavamile, noma ngisho nezintambo ze-imeyili eziqhubekayo. Lokhu kukhwabanisa nge-spear kumelela iphesenti elincane lawo wonke ama-imeyili, kodwa kuyimbangela yengxenye enkulu yokwephulwa kokuphepha okukhulu kakhulu.
ESpain, inkinga ayincane neze. Ngo-2024, kwaqoshwa amashumi ezinkulungwane zezehlakalo zokuphepha kwe-inthanethi , ukwanda okukhulu uma kuqhathaniswa nonyaka odlule, futhi ingxenye enkulu yazo yavela kuma-imeyili noma imiyalezo yokukhwabanisa. Akuyona into engaqondakali ukuthi abaphathi abaningi manje bakhomba ukuhlaselwa okukhulu noma ukwephulwa kwedatha njengenye yezingozi ezinkulu ebhizinisini labo.
Ukulinganiselwa kwabantu, izingozi, kanye nobuthakathaka
Nakuba i-AI inikeza ukuthuthuka okumangalisayo, akuyona inhlamvu yomlingo noma ikhambi elingenakugwenywa . Isadinga ukuqondiswa ngabantu, idatha yokuqeqeshwa okuhle, kanye nesu eliqinile lokuphepha kwe-inthanethi ukuze isekelwe.
Enye yobuthakathaka bomlando ekuphepheni amaphutha abantu ekucushweni kwesistimu . Izindawo ezihlanganisiwe ezinefu lomphakathi nelangasese, izinhlelo zesikhathi esidlule, kanye nezinhlelo zokusebenza ezintsha zenza ukugcina ukucushwa okuhlala njalo nokuphephile kube umsebenzi omkhulu. I-AI ingasiza ngokubona ukungahambisani, ukuphakamisa ukulungiswa, noma ngisho nokusebenzisa izinguquko ezizenzakalelayo, kodwa njalo ngaphakathi kohlaka lokulawula nokubuyekeza.
Ukukhathala komuntu kanye nokungasebenzi kahle lapho ebhekene nemisebenzi ephindaphindwayo nakho kuyinkinga. Ukulungisa ngesandla amakhulu noma izinkulungwane zama-endpoints, ukubuyekeza izexwayiso usuku nosuku, noma ukuhlola njalo ama-log ekugcineni kunciphisa ukunaka kwanoma yiliphi iqembu. Ukuzenzakalela okuhlakaniphile kuvumela le misebenzi ukuthi ilayishwe kuma-algorithms, okushiya abantu bekwazi ukusingatha ukuhumusha nezinqumo eziyinkimbinkimbi.
Ukukhathala kwezexwayiso kungenye inkinga evamile: izaziso eziningi eziqhubekayo zigcina zibangela abahlaziyi ukuthi bahlukane ngokwengqondo noma bagxile kuphela ezindabeni eziphuthumayo kakhulu, okushiya izinsongo ezingacacile kodwa eziyingozi ngokulinganayo zinganakwa. I-AI iyasiza ngokuhlela nokuhlanganisa imicimbi ehlobene futhi izibeke phambili ngokusekelwe engcupheni.
Ngaphezu kwalokho, amakhono amaqembu abantu alinganiselwe. Ukuntuleka kochwepheshe abafanelekayo bezokuphepha kwe-inthanethi kanye ne-AI/ML kusemhlabeni wonke, futhi ukuqeqesha abantu kule mikhakha kuthatha iminyaka. Amathuluzi asekelwe ku-AI avumela amaqembu amancane ukuphatha izindawo eziyinkimbinkimbi kakhulu, kodwa awaqedi isidingo sethalente labantu; ashintsha nje izinhlobo zemisebenzi eyenziwa yithalente.
Indlela i-AI kanye nokufunda komshini okusebenza ngayo ngempela ekuphepheni kwe-cyber
Kubalulekile ukuhlukanisa phakathi kwamazinga amaningana. Ngakolunye uhlangothi, kukhona ubuhlakani bokwenziwa njengomkhakha obanzi , omgomo wawo wokugcina ukuhlomisa imishini ngamakhono acishe abe ngawabantu: ukucabanga, ukuzivumelanisa nezimo, kanye nobuciko. Ukufunda komshini kuwela ngaphakathi kwalolu hlaka, futhi ukufunda okujulile, njengesigaba esithile, kufakiwe.
Empeleni, okusetshenziswa kakhulu ekuphepheni kwe-cyber namuhla ukufunda komshini (ML) , okungukuthi, amamodeli afunda kudatha yomlando ukuze enze izibikezelo kanye nokuhlukaniswa ngezigaba. Lawa mamodeli amahle kakhulu ekutholeni amaphethini, kodwa "awaqondi" ngempela umongo ngendlela umuntu angawuqonda ngayo; ukuze uthole ulwazi olujulile, ungabheka umhlahlandlela wobuchwepheshe.
Ukufunda komshini kugxila ekunembeni nasekuthuthukiseni imisebenzi ethile : uma kunikezwa isethi yedatha (isibonelo, izingodo zokuhlaselwa kwangaphambilini), ifuna indlela engcono kakhulu yokuhlukanisa phakathi kwethrafikhi evamile neyingozi. Ayizami ukuthola "isixazululo esiphelele esingcono kakhulu" senkinga yokuphepha, kodwa kunalokho ukukhulisa ukusebenza kwayo emsebenzini eqeqeshwe kuwo.
Ukufunda okujulile (DL) kuyisa lo mbono phambili ngamanethiwekhi e-neural anezingqimba eziningi akwazi ukulingisa ubudlelwano obuyinkimbinkimbi kakhulu. Ekuphepheni kwe-cyber, lawa manethiwekhi asetshenziselwa ukuhlukanisa ithrafikhi, ukuthola okungahambi kahle, ukuhlaziya ikhodi enonya, nokucubungula ulimi lwemvelo kuma-imeyili, imiyalezo, noma imibiko, yize ngezinjongo ezingokoqobo, igama elithi ML livame ukusetshenziswa kabanzi.
Inani le-ML livela ngezinhlobo ezahlukene zezinqubo: ukuhlukaniswa kwedatha (ukufaka amalebula kumafayela, ukuziphatha, noma imicimbi njengeyingozi noma enonya), ukuqoqa (ukuthola amaqembu okuziphatha okungajwayelekile ngaphandle kwamalebula angaphambilini), ukuncoma izindlela zokwenza (ukuphakamisa izinyathelo zokuphendula ngokusekelwe ezinqumweni ezedlule), noma ukubikezela kusengaphambili (ukulinganisa amathuba okuba kwenzeke isigameko noma ubuthakathaka obusetshenziswayo).
Izibonelo eziqondile ze-ML ekuphepheni kwe-cyber
Ukuze kusetshenziswe le mibono, abakhiqizi abaningi namaqembu ocwaningo babonise ukuthi ukufunda komshini (i-ML) kwandisa kanjani amakhono okuthola . Isibonelo esaziwayo yileso samaqembu okuhlaziya omhlaba wonke asebenzisa idatha evela kumanethiwekhi okuphepha asatshalaliswa emhlabeni wonke ukuqeqesha amamodeli athola izinsongo ezintsha ezithuthukisiwe, okwandisa kakhulu ukutholakala kokuhlaselwa okuqhubekayo okuthuthukisiwe (ama-APT).
Ukusetshenziswa okubanzi ukuhlukaniswa okuzenzakalelayo kanye nokuhambisana nobumfihlo bedatha . Ama-algorithms afaka ilebula yolwazi oluqukethe idatha yomuntu siqu ukuze kube lula ukuphathwa kwalo ngaphansi kwe-GDPR noma i-CCPA, okuvumela indawo esheshayo yalo lonke ulwazi oluhlobene nomsebenzisi uma esebenzisa ilungelo lakhe lokufinyelela noma lokusula.
Olunye uhlelo lokusebenza oluvamile ukudalwa kwamaphrofayili okuziphatha komsebenzisi (User Behavior Analytics) , okukuvumela ukuthi uhlukanise phakathi kwemisebenzi evamile yabasebenzi naleyo engase ibonise iziqinisekiso ezibiwe noma ukufinyelela kwangaphakathi okunonya. Izinto ezifana nokucindezela izinkinobho, izikhathi zokuxhuma, kanye nezinsiza ezifinyelelwe ziba izimpawu zokuthola abangeneleli.
Ngokufanayo, amaphrofayili okusebenza kwesistimu adalwe ukuze kuqondwe ukuthi iseva noma ikhompyutha kufanele iziphathe kanjani uma "iphilile." Uma ukusetshenziswa kwe-CPU, imemori, idiski, noma i-bandwidth kukhula ngokuzumayo ngaphandle kwencazelo ecacile, uhlelo lungabangela izexwayiso noma luhlukanise idivayisi ngenkathi ihlolwa.
Ekuvikeleni amawebhusayithi nama-API, i-ML isetshenziselwa ukuvimba ama-bot ngokusekelwe ekuziphatheni kwawo , ihlukanisa phakathi kwethrafikhi esemthethweni evela kubasebenzisi bangempela kanye namaza ezicelo ezenzakalelayo ezizama ukucindezela isevisi, ukweba okuqukethwe, noma ukuhlola iziqinisekiso ezivuvukile ngobuningi, ngisho nalapho zizama ukufihla ngemuva kwama-VPN noma ama-proxies.
I-AI ekhiqizayo, idatha, kanye namapayipi avikelekile
Ukusetshenziswa okujulile kokufunda komshini kanye ne-AI ekhiqizayo, nokho, kuletha izinselele ezinkulu maqondana nobumfihlo kanye nokuphepha kohlelo lwe-AI uqobo . Ukuqeqesha amamodeli asebenzayo kudinga idatha eningi, iningi layo elibucayi noma lomuntu siqu, elingqubuzana nezimiso ezifana "nelungelo lokukhohlwa."
Enye yezindlela zocwaningo ezithembisayo kakhulu ihilela ukukhiqiza idatha yokwenziwa elingisa idatha yangempela ngokwezibalo , okuvumela amamodeli ukuthi aqeqeshwe ngaphandle kokudalula ulwazi lomsebenzisi oluyiqiniso. Lokhu kulondoloza kangcono ubumfihlo, yize ukucwasa kanye nokuhlonza kabusha okungaqondile okungaba khona kumele kuqashwe.
Okunye okubalulekile ukuvikela lonke ipayipi le-AI: kusukela ekuqoqweni kwedatha kanye nokugcinwa kwayo kuya ekusetshenzisweni kwamamodeli ekukhiqizweni . Lokhu kuhilela ukuphathwa kwedatha okuqinile, ukubethela, ukulawula ukufinyelela, ukuqinisekiswa kwezinto eziningi, ukuhlolwa kwekhodi, kanye nokuqapha okuqhubekayo ukuthola ukuphazamiseka noma ukusetshenziswa okungagunyaziwe.
Uma imodeli ye-AI isetshenziswa kabi—ngokwesibonelo, ngedatha enobuthi— ingase yehluleke ukubona izinsongo ezithile noma ilethe ukucwasa okuyingozi ekwenzeni kwayo izinqumo . Ngakho-ke, ukuvikela ubuqotho bamamodeli kanye nedatha yabo yokuqeqesha manje sekuyingxenye ebalulekile yokuphepha kwe-cyber uqobo. Lokhu kubaluleke kakhulu ezimweni ezifana namawele edijithali.
Ngesikhathi esifanayo, ochwepheshe abaningi bafuna izinhlaka zomthetho kanye nezindinganiso ezithile ze-AI ekuphepheni kwe-inthanethi , bebhekana nakho konke kusukela ekubhekaneni namaphutha kuya ekubonakaleni okuncane okudingekayo ezinhlelweni ezenza izinqumo ezibalulekile, kanye nezidingo zokuhlola kanye nokuhlolwa njalo.
Amathuluzi okuphepha kwe-inthanethi afakiwe asebenzisa i-AI
Ngaphandle kwezigaba ezijwayelekile, kunezixazululo ezithile ezizenzele igama ngenxa yokusebenzisa kwazo kakhulu i-AI kanye ne-ML emikhakheni ehlukene yokuphepha.
Emkhakheni wasekhaya kanye nowamabhizinisi amancane, imikhiqizo ethile yenzelwe abasebenzisi be-Mac ne-Windows ngokuyinhloko , inikeza isivikelo kuma-virus, izinsongo zenethiwekhi, i-ransomware, nezinye izinhlobo ze-malware. Isici sabo sokuhlukanisa sivame ukuba sekusetshenziswa kwe-AI ukuthola izinhlobo ezintsha ngokuhlaziywa kokuziphatha, okuhlinzeka ngezeluleko ezenziwe ngezifiso ezifanelana namaphethini okusetshenziswa komsebenzisi ngamunye.
Engxenyeni yebhizinisi, abanye abakhiqizi bathuthukise amapulatifomu endabuko yamafu asebenzisa i-AI ukuthola nokuphendula i-endpoint . Lezi zixazululo zisebenzisa inzwa elula kudivayisi ngayinye, ziqoqe i-telemetry enemininingwane, bese ziyithumela epulatifomu ephakathi lapho amamodeli athuthukile ehlaziya khona ukuziphatha okungavamile, axhumanise imicimbi kumadivayisi amaningi, futhi enze izimpendulo ngokuzenzakalela.
Ezinye iziphakamiso zigxila kakhulu ekutholeni okusekelwe kunethiwekhi, zishiya indlela yendabuko esekelwe esiginesha . Ngokuhlaziywa okuqhubekayo kwethrafikhi, lezi zinhlelo zibona ukunyakaza okuseceleni, ukuphuma, kanye nemisebenzi yokuyala nokulawula, zifunda njalo ukuzivumelanisa nezinhlobo ezintsha zokuhlasela ezingabhalwanga ohlwini lwezinkomba zendabuko.
Amathuluzi asebenzisa i-AI yamahhala avele ngisho namathuluzi akhethekile ekuhlaziyeni ukukhwabanisa okungenzeka . Abasebenzisi bangalayisha isithombe-skrini, isixhumanisi, noma umbhalo osolisayo, futhi uhlelo luqhathanisa okuqukethwe kwalo nedathabheyisi enkulu yokukhwabanisa okwaziwayo, kusetshenziswa i-NLP ukuthola amaphethini akhohlisayo: ukuphuthuma okweqile, okunikezwayo okungenangqondo, izicelo zolwazi lomuntu siqu noma lwasebhange, njalo njalo.
Kuzo zonke izimo, isihluthulelo ukuthi i-AI ayigcini nje ngokusabela ezinsongweni ezaziwayo, kodwa futhi ifunda njalo emvelweni , ilungise amakhono ayo okuthola futhi inciphise ukuncika ohlwini lwabantu abamnyama noma emithethweni eqinile ephelelwa yisikhathi ngokushesha okukhulu.
Ukulungiselela ikusasa le-AI/ML kwezokuphepha kwe-cyber
Uma sibheka phambili, inhlanganisela ye-AI, i-ML, kanye ne-AI ekhiqizayo ithembisa uhlelo lokuphepha olusebenzayo noluzenzakalelayo , kodwa futhi nesimo lapho abahlaseli benamathuluzi ayinkimbinkimbi ngokulinganayo okunika amandla imikhankaso yabo.
Iminyaka ezayo kulindeleke ukuthi ibone ukuhlaselwa okunembe kakhulu nokwenziwe ngokwezifiso okunamandla okudlula izindlela eziningi zokuzivikela zendabuko, kanye nokwanda kokusetshenziswa kwe-AI ngabavikeli ukuze kutholakale, kuhlaziywe, futhi kuphendulwe ngesikhathi sangempela.
Ngenxa yalesi simo, izinhlangano zabo bonke osayizi kuzodingeka zitshale imali ekugcineni ubuchwepheshe bazo buhambisana nekusasa : ukuvuselela ingqalasizinda, ukusebenzisa amathuluzi asekelwe ku-AI aqinisekisiwe, kanye nokushiya izinhlelo eziphelelwe yisikhathi ezibeka ingozi eqhubekayo yokuxhashazwa.
Ngesikhathi esifanayo, kubalulekile ukuqaphela ukuthi i-AI kufanele ihambisane namaqembu abantu, hhayi ukuwathatha indawo . Ubuciko, ukucabanga okubucayi, ubuhlakani bebhizinisi, kanye nomthwalo wemfanelo wokuziphatha kuzohlala kungumuntu ngokusobala. Ochwepheshe bazodinga ukuqeqeshwa ukuze baqonde ukuthi la mamodeli asebenza kanjani, ukuthi ahumusha kanjani imiphumela yawo, nokuthi angawabusa kanjani ngempumelelo.
Okokugcina, ukuzivumelanisa nezimo zomthetho mayelana nedatha, ubumfihlo, kanye nokusetshenziswa kwe-AI kuzoba yingxenye ebalulekile. Ukuvuselela izinqubomgomo zangaphakathi kanye nokuhambisana nomthetho oguqukayo akuyona into ongayikhetha, ikakhulukazi emikhakheni elawulwayo lapho ukwephulwa kokuphepha kungaholela ezinhlawulweni zezigidi zamaRandi kanye nomonakalo wedumela okunzima ukulungisa.
Konke kukhomba ikusasa lapho ukubambisana phakathi kwabantu nemishini kuzoba yitshe lesisekelo sokuzivikela kwedijithali : ukuphatha ukuqapha okuqhubekayo, ukuhlaziywa kwedatha okukhulu, kanye nempendulo yokuqala ezenzakalelayo, kanye namaqembu okuphepha kwe-inthanethi enza izinqumo zamasu, ukucwenga amamodeli, kanye nokuklama amasu omhlaba wonke okugcina izinhlelo ziphephile endaweni esongelayo ehlala ishintsha.
