OpenAI yabujije konti z’Uburusiya zifitanye isano n’ibitekerezo by’impimbano bya Isiraheli, nk'uko Hardware ya Tom ibitangaza

Byagenze bite

Tom's Hardware ivuga ko OpenAI yabujije ihuriro rya konti ya ChatGPT ku ya 25 Kanama nyuma yo kubahuza n’ikigo mpuzamahanga cya Burke, ikigo cy’ibitekerezo cya Isiraheli. OpenAI yavuze ko abayikoresheje bakoresheje VPN kugira ngo bagere kuri serivisi zayo mu Burusiya kandi bakoresha ChatGPT kugira ngo batange inyandiko zamamaza n'ibisubizo, atari ingingo zishingiye ku rubuga.

Icyuma cya Tom cyatangaje ko OpenAI yabujije ihuriro rya konti ya ChatGPT ku ya 25 Kanama maze ikurikirana icyo gikorwa mu Burusiya. Konti zahujwe n’ikigo mpuzamahanga cya Burke, cyigaragaje nkikigo cy’ibitekerezo cya Isiraheli. Nk’uko iyi raporo ibigaragaza, uru rubuga rwashyize ahagaragara ubushakashatsi bwa politiki y’abimukira bwitirirwa umwarimu wo muri Ositaraliya w’ubumenyi bw’ibiribwa, bukora “indangagaciro y’ubusugire” bwerekanaga Uburusiya hejuru y’ibihugu binenga intambara yarwo, kandi bukakira impapuro z’amasomo zandukuwe n’abanditsi babo ba mbere. Inkomoko ntisobanura abayikora cyangwa ngo bavuge ko bakoraga mu izina rya guverinoma y’Uburusiya.

Uruhare rwavuzwe na ChatGPT rwakoraga cyane. OpenAI yabwiye ibyuma bya Tom ko abakoresha bakoresha chatbot kugirango bandike inyandiko n'ibisubizo biteza imbere ibicuruzwa hirya no hino muri Substack, Telegram, X, Facebook na LinkedIn. Bakoresheje kandi kugirango bakureho amagambo ashobora guhishura abanditsi bakomoka mu kirusiya. OpenAI yavuze ko ingingo, raporo z’igihugu ndetse n’ibipimo ubwabyo bitakozwe hamwe n’icyitegererezo cyayo. Raporo isobanura ingingo imwe yakoresheje “ihuriro rya Svetofor,” ubusobanuro bw'ijambo ry'ikirusiya rivuga itara ry'umuhanda, nk'ikimenyetso kigaragaza ko ibikoresho bitari AI bigifite ibimenyetso by'indimi bijyanye n'inkomoko yabyo.

Tom's Hardware ivuga ko OpenAI yakusanyije ingingo 36 zo mu kigo mpuzamahanga cya Burke isanga 34 zandukuwe n’abize nyabo. Bamwe bahawe nabi abantu bakomeye barimo Francis Fukuyama na Noam Chomsky. Ikinyamakuru kivuga kandi ko OpenAI yashyize mu bikorwa ubukangurambaga ku mpera yo mu cyiciro cya gatatu ku gipimo cya Brookings. Raporo ivuga ko buri muntu ku giti cye yakiriye umubare muto wo kureba, mu gihe imiyoboro ya Telegramu y’igikorwa yari ifite abayoboke bari hagati ya 10,000 na 20.000 buri umwe. Iyi mibare isobanura ibikoresho OpenAI yasuzumwe, ntabwo igerwaho cyangwa ingaruka zikorwa mugari.

Ibisobanuro birambuye: tomshardware.com ↗

Impamvu ari ngombwa

Uru rubanza rwerekana uburyo AI ibyara inyungu ishobora gushyigikira ibikorwa byoguhindura binyuze mubusemuzi, gutunganya, kubateze amatwi no gukwirakwiza nubwo ibikoresho nyamukuru byandukuwe cyangwa byanditswe nabantu. Icyuma cya Tom's Hardware kivuga ko ubukangurambaga butageze aho bugarukira, ariko ibikorwa remezo byagutse kuruta ababyumva.

The case is significant because AI was used as an efficiency and distribution layer in a political influence operation rather than as the original author of its core content. Tom’s Hardware says OpenAI’s Intelligence & Investigations team has repeatedly found operators attaching AI tools to existing workflows for tasks such as drafting, editing and audience development. That distinction matters: removing generated articles from the picture does not remove the ability of a chatbot to help an operator produce more persuasive or more localized outreach at greater speed.

The reported campaign also illustrates how fabricated institutional identities can make copied material appear authoritative. A think tank format, named researchers, country reports and an index can give readers multiple signals of legitimacy even when the underlying work is misattributed or taken without permission. AI-assisted language polishing may further reduce obvious signs that material originated elsewhere. The source provides no evidence that the campaign achieved substantial public persuasion, but the combination of false branding, copied scholarship and cross-platform promotion creates a practical verification problem for readers and publishers.

The enforcement action also raises questions about the scope and transparency of platform disruption. Tom’s Hardware reports that this was OpenAI’s sixth public threat report since early 2024 and that the company says it has disrupted more than 40 malicious networks. OpenAI characterized the operation’s importance as lying less in its audience than in the infrastructure it built. That assessment is the company’s view, however. The source does not independently verify the account links, the scale of the network, the reported follower counts or whether the banned users operated additional accounts and websites.

Ibyo kureba

Inkomoko ntabwo yigenga yigenga ibiranga OpenAI cyangwa gushiraho uwakoresheje konti. Ibindi bisobanuro bigomba gusuzuma impapuro zandukuwe, hamagara abize ibyangijwe no gusuzuma niba ikigo mpuzamahanga cya Burke gikomeje gukora kurubuga rusange.

The first priority is independent verification of the International Burke Institute’s material. The source identifies copied work and examples of questionable attribution, but it does not include responses from the academics whose work was allegedly lifted. Follow-up reporting should establish which papers were copied, whether they were altered, whether the named scholars were contacted and how the website represented their work. It should also distinguish confirmed copying from OpenAI’s broader characterization of the operation.

The status of the associated distribution channels remains unclear. Tom’s Hardware reports activity across Substack, Telegram, X, Facebook and LinkedIn, but does not say whether those accounts or channels were removed, labeled or left online after the ChatGPT accounts were banned. The report also does not establish how many people saw the campaign outside the sampled posts or whether any material was republished by unrelated users. Those unknowns are important for assessing public impact rather than simply measuring the size of the infrastructure.

Hanyuma, reba ibisobanuro birambuye kubyerekeranye na OpenAI hamwe nuburyo bwo kubahiriza. Amakuru avuga ko iyi sosiyete yakurikiranye iryo tsinda mu Burusiya kandi ko abayikoresheje bakoresheje VPN kubera ko serivisi za OpenAI zahagaritswe aho, ariko ntisobanura ibimenyetso bya tekiniki, umubare wa konti cyangwa igihe nyacyo cy’ibikorwa byabo. Ntabwo kandi ivuga niba abategetsi b’Uburusiya, ikigo cy’ibitekerezo gikekwa ko ari amashami cyangwa urubuga rutwara ibikoresho byamamaza byashubije. Kugeza ubwo ibyo bibazo bisubijwe, amakuru yatangajwe agomba kuguma yitirirwa ibyuma bya Tom na OpenAI aho gufatwa nkibigenga.

Ibijyanye nuyobora & ibibazo