Ukuhlangana Kokukhubazeka Kosuku Olungenalo Nentuthuko ye-AI: Ukukhathazeka Okukhulayo
Imboni yezobuchwepheshe izulazula endaweni eyinkimbinkimbi lapho ubungozi bezinsuku eziyiziro kanye nentuthuko esheshayo ye-AI hlangana khona, okudala isivunguvungu esifanele ochwepheshe bezokuphepha okufanele basilungise. Ukuze uqonde le nkinga, kubalulekile ukuchaza amagama abalulekile: ukuba sengozini kwezinsuku eziyiziro kubhekisela emaphutheni okuvikela angaziwa ngaphambili angaxhashazwa abahlaseli, kuyilapho intuthuko ye-AI ihilela ukuthuthukiswa kwamasistimu obuhlakani bokwenziwa angakwazi ukwenza imisebenzi ngokuzenzakalelayo. Njengoba kubikiwe [1], inani lobungozi bezinsuku eziyiziro elixhashazwayo endle liyakhula, ngezibonelo eziphawulekayo ezihlanganisa ubungozi obubalulekile obutholwe emikhiqizweni ye-Check Point kanye nama-iPhone nama-iPads e-Apple. Ngaleso sikhathi, indawo ye-AI ibona ukuqanjwa okusha okubalulekile, okuhlanganisa ukwethulwa kwe-GPT-5.6 Sol kanye ne-Grok 4.5, kanye nokutholwa kwe-OpenAI kwe-Ona ye-Codex [2].
Ukwanda kwakamuva kobungozi bezinsuku eziyiziro kugqamisa usongo olukhulayo lokuhlaselwa kwezinsuku eziyize, okusebenzisa ubungozi obungaziwa ngaphambilini, okushiya amaqembu onogada engenasikhathi esincane sokuphendula. Isibonelo, ukuba sengozini kwe-Check Point yosuku oluyiziro, i-CVE-2026-16232, kube nemiphumela enzima, ehlanganisa ukonakala kwedatha ebucayi kanye nokuphazamiseka kwezinhlelo ezibucayi [12]. Iqiniso lokuthi i-CISA yengeze lobu bungozi kukhathalogi yayo eyaziwa ngokuthi yi- Known Exploited Vulnerabilities (KEV) [13] igcizelela ubukhali bosongo. Ukubonisa umthelela wokuhlasela kosuku oluyiziro, cabangela indaba yokwephulwa kwe-Equifax, lapho ukuba sengozini kwezinsuku eziyiziro kusofthiwe ye-Apache Struts kuholele ekuvezweni kwedatha ebucayi kubantu abangaphezu kwezigidi ezingu-147.
Ijubane elisheshayo le-AI emisha inkemba esika nhlangothi zombili. Ngakolunye uhlangothi, i-AI inamandla okuguqula izimboni futhi iguqule indlela esiphila ngayo nesisebenza ngayo. Ngakolunye uhlangothi, ukuthembela okwandayo ezinhlelweni ze-AI kudala ubungozi obusha kanye nezindawo zokuhlasela. Njengoba izinhlelo ze-AI ziya zigcwele, ingozi yokuhlaselwa okunamandla e-AI iyakhula. Ngokuka- [14], konke ukwephulwa kwezokuphepha kwe-ejenti ye-AI okuqinisekisiwe, ukuxhashazwa kwe-CVE, nokuba sengozini kuyalandelelwa, okugqamisa isidingo sokuqapha lapho kubhekwa izinsongo ezivelayo. Isibonelo, ukuhlasela kobugebengu bokweba imininingwane ebucayi okuxhaswe yi-AI kungaba yinkimbinkimbi kakhulu, kusetshenziswa ama-algorithms okufunda komshini ukuze kugwenywe ukutholwa nokukhohlisa abasebenzisi ukuthi bakhiphe ulwazi olubucayi.
Ukuxhumana phakathi kokuba sengozini kwezinsuku eziyiziro kanye nentuthuko ye-AI kulele eqinisweni lokuthi amasistimu e-AI awagonyiwe ekuhlaselweni kwezinsuku eziyiziro. Eqinisweni, amasistimu e-AI angasetshenziswa ukuqalisa ukuhlasela okuyinkimbinkimbi okuxhaphaza ubungozi bosuku oluyiziro. I-AegisAI, inkampani eyasungulwa yizikhulu zezokuphepha ze-Google zangaphambili, isungule isisombululo esinamandla e-AI ukumisa ukuhlaselwa kobugebengu bokweba imininingwane ebucayi obuqhutshwa yi-AI [9]. Lokhu kugqamisa igeyimu yekati negundane phakathi kokuhlasela nokuvikela okunamandla e-AI. Ukuze uqhubeke ucacisa leli phuzu, cabanga ngendaba yokuhlasela kwe-NotPetya ransomware, eyasebenzisa amasu obunjiniyela bomphakathi anikwe amandla yi-AI ukuze isabalalise futhi igweme ukutholwa.
Ukuze ubeke ukuphepha eqhulwini lapho kubhekene nokuba sengozini kwezinsuku eziyiziro kanye nentuthuko ye-AI, ochwepheshe bezokuphepha kufanele bahlale benolwazi mayelana nezinsongo zakamuva namathrendi. Lokhu kubandakanya imithombo yokuqapha efana ne-[4] kanye ne-[6] yezindaba zakamuva nezibuyekezo ngobungozi bezinsuku eziyiziro kanye nentuthuko ye-AI. Ukwengeza, amathimba ezokuphepha kumele abeke phambili amasistimu wokuchibiyela nokuvuselela, kanye nokusebenzisa izindlela zokuphepha eziqinile ukuze kuvinjelwe ukuhlaselwa kwezinsuku eziyiziro. Isibonelo, ukusebenzisa inqubo eqinile yokulawulwa kweziqephu kungasiza ekwehliseni ingcuphe yokuhlaselwa kosuku oluyiziro, kuyilapho ukusebenzisa amathuluzi okuvikela anamandla e-AI kungasiza ekutholeni futhi kuphendule izinsongo ezivelayo.
Ikusasa lokuphepha lizololongwa ukuphambana kobungozi bezinsuku eziyiziro kanye nentuthuko ye-AI. Njengoba izinhlelo ze-AI ziya zigcwele, ingozi yokuhlaselwa okunamandla e-AI izokhula. Kodwa-ke, i-AI ingase futhi isetshenziselwe ukuthuthukisa ukuphepha, njengokuthuthukiswa kwezixazululo ezisebenza ngamandla e-AI ukumisa ukuhlasela kobugebengu bokweba imininingwane ebucayi obuqhutshwa yi-AI. Ngokuka- [3], ngoJulayi 2026 kwasungulwa i-AI engama-27 embozwe ngo-ThursdAI, eholwa yi-OpenAI, i-Anthropic, i-Google DeepMind, i-Meta AI kuwo wonke ama-ejenti, amamodeli wemingcele, umthombo ovulekile, nokubhala ikhodi. Lokhu kugqamisa ijubane elisheshayo lokuqamba izinto ezintsha endaweni ye-AI kanye nesidingo sokuthi ochwepheshe bezokuphepha bahlale ngaphambi kwejika.
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/