- I-Radware ithole ubungozi ku-ChatGPT Deep Research engakhipha idatha ye-akhawunti ye-Gmail.
- Ukuhlasela kusetshenziswe umjovo wokwaziswa ongaqondile onemiyalo ye-HTML efihliwe futhi kwaqhutshwa nengqalasizinda ye-OpenAI.
- I-OpenAI isivele yalinciphisa iphutha; abukho ubufakazi bomphakathi bokuxhashazwa kwangempela.
- Kunconywa ukuthi ubuyekeze futhi uhoxise izimvume ku-Google futhi ukhawulele ukufinyelela kwabenzeli be-AI kuma-imeyili namadokhumenti.
Ucwaningo lwakamuva selutholile imbobo yokuphepha kumenzeli Wocwaningo Olujulile we-ChatGPT lokho, ngaphansi kwezimo ezithile, kungase kube lula ukukhishwa kolwazi oluvela kuma-imeyili asingathwe ku-GmailUkutholwa kugqamisa ubungozi bokuxhuma abasizi be-AI kumabhokisi okungenayo namanye amasevisi aqukethe idatha ebucayi.
Inkampani ye-cybersecurity I-Radware ibike lolu daba ku-OpenAI, futhi umthengisi wayinciphisa ngasekupheleni kwehlobo ngaphambi kokuthi ibe nolwazi lomphakathi.. Nakuba isimo sokuxhashazwa sasilinganiselwe futhi abukho ubufakazi bokuhlukunyezwa emhlabeni wangempela, indlela esetshenziswa amaqabunga isifundo esibalulekile kubasebenzisi namabhizinisi.
Kwenzekeni ngedatha ye-ChatGPT ne-Gmail?
Ucwaningo Olujulile luyi-ejenti ye-ChatGPT okuqondiswe ophenyweni olunezinyathelo eziningi okungase, uma umsebenzisi eyigunyaza, axhumane imithombo eyimfihlo njenge-Gmail ukukhiqiza imibiko. Iphutha livule umnyango ukuze umhlaseli alungiselele umlayezo othile, futhi isistimu, lapho ihlaziya ibhokisi lokungenayo, ingalandela imiyalo engadingeki.
Ubungozi bangempela buncike kumuntu ocela i-ChatGPT ukuthi yenze uphenyo oluthile nge-imeyili yakhe kanye nokuthi udaba ifanise nokuqukethwe kwe-imeyili enonya. Noma kunjalo, i-vector ikhombisa ukuthi i-ejenti ye-AI ingaba kanjani ucezu olusiza ukuvuza kwedatha.
Phakathi kolwazi okungenzeka luthinteke lungavela amagama, amakheli noma enye idatha yomuntu siqu ekhona emilayezweni ecutshungulwe umenzeli. Lokhu bekungekona ukufinyelela okuvulekile ku-akhawunti, kodwa kunalokho ukuhlunga okubekwe wumsebenzi owabelwe umsizi.
Isici esibucayi kakhulu ukuthi umsebenzi uqale kusukela ku- Ingqalasizinda yefu ye-OpenAI, okwenze kwaba nzima ngezivikelo zendabuko ukuthola ukuziphatha okuxakile njengoba kungaveli kudivayisi yomsebenzisi.
I-ShadowLeak: Umjovo Osheshayo Owenze Kwenzeka
I-Radware ibize inqubo ethi ShadowLeak futhi uyifake ku-a umjovo osheshayo ongaqondile: imiyalelo efihliwe ngaphakathi kokuqukethwe i-ejenti ekuhlaziyayo, ekwazi ukuba nomthelela ekuziphatheni kwayo ngaphandle kokuqaphela komsebenzisi.
Umhlaseli uthumele i-imeyili nge imiyalelo ye-HTML efihliwe ngokusebenzisa amaqhinga afana namafonti amancane noma umbhalo omhlophe kungemuva elimhlophe. Uma uthi nhlá I-imeyili ibonakala ingenabungozi, kodwa yayihlanganisa imiyalelo yokusesha ibhokisi lokungenayo ukuze uthole idatha ethile..
Lapho umsebenzisi ecela Ucwaningo Olujulile ukuthi lusebenze ku-imeyili yakhe, i-ejenti yafunda leyo miyalo engabonakali futhi iqhubekile nokukhipha nokuthumela idatha kuwebhusayithi elawulwa umhlaseliEzivivinyweni, abacwaningi baze bafinyelela ekubhaleni imininingwane ku-Base64 ukuze ibonakale njengesinyathelo sokuphepha.
Izithiyo ebezidinga imvume ecacile ukuze uvule izixhumanisi nazo zingagwenywa ngokucela amathuluzi okuzula omenzeli, aqondise ukuhlunga ezizindeni zangaphandle ngaphansi kokulawula umhlaseli.
Ezindaweni ezilawulwayo, Amaqembu e-Radware aphawule izinga eliphezulu kakhulu lokusebenza ngempumelelo, okubonisa ukuthi inhlanganisela yokufinyelela i-imeyili nokuzimela kwe-ejenti kungaba ukukholisa imodeli uma imiyalelo eshumekiwe ingahlungiwe kahle.
Kungani ingazange ibonwe ngabavikeli
Ukuxhumana bekuvela kumaseva athembekile, ngakho-ke amasistimu ezinkampani abone ithrafikhi esemthethweni evela kusevisi enedumela elihle. Le mininingwane iphendule ukuvuza kwaba a indawo eyimpumputhe yezixazululo eziningi ukuqapha.
Ngaphezu kwalokho, isisulu bekungadingeki ukuthi sichofoze noma senze noma yini eqondile: uvele wacela umenzeli ukusesha okuhlobene nesihloko se-imeyili esilungiswe umhlaseli, into eyenza ukuqondisa. buthule futhi kunzima ukulandelela.
Abacwaningi bayakugcizelela lokho Sibhekene nohlobo olusha lokusongela lapho i-ejenti ye-AI ngokwayo isebenza njenge-vector. Noma nomthelela osebenzayo okhawulelwe, icala liyasiphoqa ukuthi sibuyekeze ukuthi sinikeza kanjani izimvume kumathuluzi azenzakalelayo.
Ukulungiswa kwephutha nezincomo ezisebenzayo
I-OpenAI isebenzise ukuncishiswa kulandela isaziso se-Radware futhi izwakalise ukubonga kwayo ngobufakazi obuphikisanayo, igcizelela ukuthi ihlale iqinisa izivikelo zayo. Kuze kube manje, umhlinzeki ufuna lokho abukho ubufakazi bokuxhashazwa yale vector.
I-Deep Research iwumenzeli ozikhethela wona ongaxhuma kuphela ku-Gmail ngemvume ecacile yomsebenzisi. Ngaphambi kokuxhuma amabhokisi okungenayo noma amadokhumenti kumsizi, Kunconywa ukuhlola ububanzi bangempela bezimvume futhi unciphise ukufinyelela kulokho okudingeka kakhulu..
Uma uxhume izinsiza ze-Google, buyekeza futhi ulungise ukufinyelela Kulula:
- Iya ku-myaccount.google.com/security ukuvula iphaneli yokuvikela.
- Esigabeni sokuxhumana, chofoza okuthi Buka konke ukuxhumana.
- Khomba i-ChatGPT noma ezinye izinhlelo zokusebenza ongazazi futhi uhoxise izimvume..
- Susa ukufinyelela okungadingekile futhi uphinde unikeze kuphela okudingeka kakhulu. okubalulekile.
Kubasebenzisi namabhizinisi, Kusemqoka ukuhlanganisa umqondo ojwayelekile kanye nezinyathelo zobuchwepheshe: gcina yonke into isesikhathini, sebenzisa umgomo wokungabi namalungelo amancane kuma-ejenti nezixhumi., futhi igade umsebenzi wamathuluzi ngokufinyelela kudatha ebucayi.
Ezindaweni zezinkampani, ochwepheshe batusa ukufaka izilawuli ezengeziwe zama-ejenti we-AI futhi, uma Ucwaningo Olujulile noma izinsizakalo ezifanayo zisetshenziswa, khawulela amakhono njengokuvula izixhumanisi noma ukuthumela idatha ezizindeni ezingaqinisekisiwe.
Ucwaningo lwe-Radware kanye nokunciphisa okusheshayo kwe-OpenAI kushiya isifundo esicacile: ukuxhuma abasizi ku-Gmail kunikeza izinzuzo, kodwa izidingo zokuphepha hlola izimvume, qapha ukuziphatha futhi ucabange ukuthi umjovo wemiyalo uzoqhubeka nokuhlola ama-agent e-AI.
Ngingumshisekeli wezobuchwepheshe oguqule izintshisekelo zakhe "ze-geek" zaba umsebenzi. Ngichithe iminyaka engaphezu kwengu-10 yempilo yami ngisebenzisa ubuchwepheshe obusezingeni eliphezulu kanye nokukitaza ngazo zonke izinhlobo zezinhlelo ngenxa yelukuluku lokufuna ukwazi. Manje sengiqeqeshelwe ubuchwepheshe be-computer nemidlalo yama-video. Lokhu kungenxa yokuthi sekuphele iminyaka engaphezu kwengu-5 ngisebenza ngokubhalela amawebhusayithi ahlukahlukene ezobuchwepheshe nemidlalo yevidiyo, ngenza izindatshana ezifuna ukukunikeza imininingwane oyidingayo ngolimi oluqondakala yiwo wonke umuntu.
Uma unemibuzo, ulwazi lwami lusukela kuyo yonke into ehlobene nesistimu yokusebenza ye-Windows kanye ne-Android yomakhalekhukhwini. Futhi ukuzibophezela kwami kuwe, ngihlala ngizimisele ukuchitha amaminithi ambalwa futhi ngikusize uxazulule noma yimiphi imibuzo ongase ube nayo kulo mhlaba we-inthanethi.