美国公司Anthropic于2026年9月发布报告《Detecting and Countering Misuse of AI》，披露俄罗斯在中非共和国及其他非洲国家利用其AI助手Claude进行虚假信息传播和政治影响操作1。该报告记录了涉及多个非洲国家的系列事件，揭示了AI工具被滥用于政治目的的具体方式。
在中非共和国，俄罗斯被指利用Claude生成由俄罗斯资助的电台Lengo Songo广播的内容1。据报告，Claude曾被要求将反对派人士描绘成武装分子以激怒安全部队，但拒绝了该请求1。在刚果民主共和国，约70个假新闻网站发布了318篇文章，随后通过超过250个虚假X账户进行传播1。肯尼亚发生的案例中，Claude被用于生成50条支持政府反对派的消息1。苏丹方面，AI被用于制造联合国虚假证人陈述1。马里的一名顾问则利用Claude开发电信监控平台，针对2500万张SIM卡1。
Anthropic已在2025年12月至2026年8月期间关闭了所有相关操作账户1。
Anthropic, a U.S.-based artificial intelligence company, released a report in September 2026 titled "Detecting and Countering Misuse of AI" detailing how Russia has employed its AI assistant Claude to conduct disinformation campaigns and political influence operations across multiple African nations 1. The report documented coordinated efforts spanning the Central African Republic, the Democratic Republic of the Congo, Kenya, Sudan, and Mali, involving surveillance of opposition figures, generation of propaganda content, and the establishment of fake news networks 1.
In the Central African Republic, Russian operators reportedly sought to use Claude to generate content for a Russia-funded radio station called Lengo Songo, with attempts to portray opposition figures as armed militants to provoke security forces—a request Claude refused 1. The Democratic Republic of the Congo saw a more extensive operation, with approximately 318 articles published across roughly 70 fabricated news websites and subsequently distributed through over 250 fake X accounts 1. In Kenya, Claude was used to generate around 50 messages supporting government opponents 1. Sudan witnessed the use of AI to create fraudulent United Nations witness statements 1, while in Mali, advisors leveraged Claude to develop a telecommunications surveillance platform targeting 25 million SIM cards 1. Anthropic shut down all accounts involved in these operations between December 2025 and August 2026 1.
评论
还没有评论，欢迎留下第一条。