Ukudibana koBusichenene boSuku lweZiro kunye nokuPhucula i-AI: Inkxalabo ekhulayo
Ishishini letekhnoloji lijonga kumhlaba ontsonkothileyo apho ubuthathaka bosuku-zero kunye nokuqhubela phambili ngokukhawuleza kwe-AI kuhlangana, kudala isaqhwithi esigqibeleleyo ekufuneka iingcali zokhuseleko zijongane. Ukuze uqonde lo mba, kubalulekile ukuchaza amagama angundoqo: ubuthathaka beentsuku zero bubhekisa kwiziphene ezingaziwayo zokhuseleko ezinokuthi zixhatshazwe ngabahlaseli, ngelixa ukuqhubela phambili kwe-AI kubandakanya uphuhliso lweenkqubo zobukrelekrele ezenziweyo ezinokwenza imisebenzi ngokuzimeleyo. Njengoko kuxeliwe [1], inani lobuthathaka beentsuku zero ezixhatshazwayo endle liyanda, kunye nemizekelo ephawulekayo kubandakanya ubuthathaka obubalulekileyo obufunyenwe kwiimveliso zeCheck Point kunye nee-iPhones kunye nee-iPads ze-Apple. Okwangoku, indawo ye-AI ibona izinto ezintsha, kubandakanywa ukuqaliswa kwe-GPT-5.6 Sol kunye ne-Grok 4.5, kunye nokufumana i-OpenAI ye-Ona yeCodex [2].
Ukuqhambuka kwamva nje kobuthathaka beentsuku zero kuqaqambisa isoyikiso esikhulayo sohlaselo lwemini engu-0, olusebenzisa ubuthathaka obungaziwa ngaphambili, beshiya amaqela okhuseleko enexesha elincinci lokuphendula. Ngokomzekelo, i-Check Point ye-zero-day vulnerability, i-CVE-2026-16232, ibe nemiphumo emibi, kubandakanywa ukuthotyelwa kwedatha ebuthathaka kunye nokuphazamiseka kweenkqubo ezibalulekileyo [12]. Inyani yokuba i-CISA yongeze obu buthathaka kwiCatalog yayo eyaziwa ngokuba yi- Known Exploited Vulnerabilities (KEV) [13] igxininisa ubuzaza bosoyikiso. Ukubonisa ifuthe lohlaselo lweentsuku zero, qwalasela imeko yokwaphulwa kwe-Equifax, apho ukuba sesichengeni kosuku lwe-zero kwisoftware ye-Apache Struts kukhokelele ekuvezweni kwedatha ebuthathaka kubantu abangaphezu kwezigidi ezili-147.
Isantya esikhawulezayo sokuqaliswa kwe-AI sikrele elintlangothi-mbini. Kwelinye icala, i-AI inamandla okuguqula amashishini kunye nokuguqula indlela esiphila ngayo nesisebenza ngayo. Kwelinye icala, ukuthembela okwandayo kwiinkqubo ze-AI kudala ubuthathaka obutsha kunye nokuhlaselwa komphezulu. Njengoko iinkqubo ze-AI zisasazeka ngakumbi, umngcipheko wokuhlaselwa kwe-AI ukhula. Ngokuka [14], yonke into eqinisekisiweyo yokwaphulwa kokhuseleko lwearhente ye-AI, ukuxhaphazwa kwe-CVE, kunye nokuba sesichengeni kuyalandelelwa, kuqaqambisa imfuno yokuphaphela ebusweni bezoyikiso ezivelayo. Umzekelo, uhlaselo oluqhutywa yi-AI lunobunkunkqele obuphezulu, kusetyenziswa ii-algorithms zokufunda koomatshini ukuphepha ukubhaqwa kunye nokukhohlisa abasebenzisi ukuba baveze ulwazi olubuthathaka.
Uqhagamshelo phakathi kobuthathaka beentsuku zero kunye nokuqhubela phambili kwe-AI kukwinto yokuba iinkqubo ze-AI azikhuselekanga kuhlaselo lwemini engu-0. Ngapha koko, iinkqubo ze-AI zinokusetyenziselwa ukuqalisa uhlaselo oluntsonkothileyo olusebenzisa ubuthathaka bosuku lwe-0. I-AegisAI, inkampani eyasekwa ngabaphathi bezokhuseleko bangaphambili bakaGoogle, iphuhlise isisombululo esine-AI sokumisa uhlaselo lomkhonto oluqhutywa yi-AI [9]. Oku kuqaqambisa umdlalo wekati kunye nemouse phakathi kohlaselo olunamandla lwe-AI kunye nokhuselo. Ukucacisa ngakumbi le ngongoma, qwalasela imeko yohlaselo lwe-NotPetya ransomware, olusebenzisa ubuchule bobunjineli bezentlalo obuxhaswe yi-AI ukusabalalisa kunye nokuphepha ukubhaqwa.
Ukubeka phambili ukhuseleko ebusweni bobuthathaka beentsuku zero kunye nokuqhubela phambili kwe-AI, iingcali zokhuseleko kufuneka zihlale zinolwazi malunga nezoyikiso zamva nje kunye neendlela. Oku kubandakanya imithombo yokubeka iliso efana ne- [4] kunye ne- [6] yeendaba zamva nje kunye nohlaziyo malunga nobuthathaka beentsuku zero kunye nokuqhubela phambili kwe-AI. Ukongezelela, amaqela okhuseleko kufuneka abeke phambili ukubeka phambili kunye nokuhlaziya iinkqubo, kunye nokuphumeza amanyathelo okhuseleko aqinileyo ukukhusela ukuhlaselwa kweentsuku zero. Ngokomzekelo, ukuphumeza inkqubo yokulawula i-patch eqinile kunokuncedisa ukunciphisa umngcipheko wokuhlaselwa kwe-zero-day, ngelixa ukusebenzisa izixhobo zokhuseleko ezinikwe amandla e-AI kunokunceda ukufumanisa nokuphendula kwizoyikiso ezivelayo.
Ikamva lokhuseleko liza kubunjwa kukudityaniswa kobuthathaka beentsuku zero kunye nokuqhubela phambili kwe-AI. Njengoko iinkqubo ze-AI zisasazeka ngakumbi, umngcipheko wokuhlaselwa kwe-AI uya kukhula. Nangona kunjalo, i-AI ingasetyenziselwa ukomeleza ukhuseleko, njengophuhliso lwezisombululo ze-AI zokunqanda uhlaselo lomkhonto oluqhutywa yi-AI. Ngokuka [3], kaJulayi ka-2026 yabona i-27 ye-AI yasungulwa ngo-ThursdAI, ikhokelwa yi-OpenAI, i-Anthropic, i-Google DeepMind, i-Meta AI kwii-arhente, iimodeli zemida, umthombo ovulekileyo, kunye nekhowudi. Oku kugxininisa isantya esikhawulezayo sokwenziwa kwezinto ezintsha kwindawo ye-AI kunye nemfuneko yokuba iingcali zokhuseleko zihlale phambi kwegophe.
Imithombo
- https://techcrunch.com/2026/07/
- https://aitoolsrecap.com/Blog/AINewsJuly2026.aspx
- https://thursdai.news/releases/2026-07
- https://techcrunch.com/
- https://aiweekly.co/ai-news-today
- https://www.timesofai.com/news/ai/
- https://thirdruntime.com/
- https://www.secondtalent.com/news/tech/
- https://techcrunch.com/2026/
- https://nvd.nist.gov/vuln/detail/CVE-2026-58644
- https://www.cvefind.com/
- https://www.securityweek.com/new-check-point-zero-day-vulnerability-exploited-in-the-wild/
- https://www.cisa.gov/news-events/alerts/2026/07/22/cisa-adds-two-known-exploited-vulnerabilities-catalog
- https://axis-intelligence.com/ai-agent-security-incident-tracker/
- https://securityonline.info/