AI tools become part of cross-border influence campaigns
国境を越える情報工作にAIツール、OpenAIが報告テクノロジー
- メディア
- NPR Business
- 国
- 🇺🇸 United States
日本語の要約
英語音声
Voice A
英語本文
OpenAI says it removed accounts linked to Russian and Iranian influence operations that used its artificial intelligence tools. NPR reported the company's findings, which describe efforts to spread misleading narratives in several countries. The accounts did not all use AI in the same way. Some used it mainly to report their activities to supervisors, while others used it to draft material for publication.
According to OpenAI, Russian operatives used ChatGPT largely to prepare internal reports. Some of those reports claimed successes that the company could not attribute to the operators. The Iranian-linked activity involved creating and polishing content in Persian and English, including articles submitted under invented author identities. These are the company's assessments of the activity, rather than proof that every claim made by the operators was true.
The company said the Iranian accounts created seven fictitious authors and published nearly 100 articles through small outlets around the world. NPR also described how material reached established media and how one outlet removed author pages afterward. The findings highlight a verification problem for publishers: checking whether writing sounds convincing is different from checking who produced it and whether its claims are supported.