Breaking
ITMexico: Sheinbaum sends 800 soldiers to Sinaloa after the killing of a US citizenUSU.S. Military Acknowledges First Deployment of Space WeaponBRSTF session is suspended after argument between ministers over the actions of the Federal PoliceRUAn air raid alert has been declared throughout UkraineARWidespread international condemnation of the Houthi attacks on Saudi Arabia and Saudi-Egyptian discussions to enhance regional securityINUS reveals it has placed weapons in spaceFRRC Lens: Dino Toppmöller leaves his post as coachFRMaizières-lès-Metz: a woman in police custody after the desecration of nearly 100 gravesRUA man was hospitalized after an Ukrainian Armed Forces drone attack in ShebekinoGLOBALHealth Secretary Yvette Cooper to introduce CCTV on baby wards following Lucy Letby inquiryITMexico: Sheinbaum sends 800 soldiers to Sinaloa after the killing of a US citizenUSU.S. Military Acknowledges First Deployment of Space WeaponBRSTF session is suspended after argument between ministers over the actions of the Federal PoliceRUAn air raid alert has been declared throughout UkraineARWidespread international condemnation of the Houthi attacks on Saudi Arabia and Saudi-Egyptian discussions to enhance regional securityINUS reveals it has placed weapons in spaceFRRC Lens: Dino Toppmöller leaves his post as coachFRMaizières-lès-Metz: a woman in police custody after the desecration of nearly 100 gravesRUA man was hospitalized after an Ukrainian Armed Forces drone attack in ShebekinoGLOBALHealth Secretary Yvette Cooper to introduce CCTV on baby wards following Lucy Letby inquiry
Back中国国安部聚焦AI风险:政治安全列为首要挑战
中国国安部聚焦AI风险:政治安全列为首要挑战
Tech
德国之声1 hour agoTech4 min readChinaView translation

中国国安部聚焦AI风险:政治安全列为首要挑战

中国国安部部长陈一新撰文指出人工智能带来的危险挑战,分析人士指出这反映出威权政权对维系权力的执迷。

Quick Look

中国国安部部长陈一新在最新文章中列举了人工智能带来的政治安全、黑客攻击及窃密等危险挑战。墨卡托中国研究中心专家指出,这体现了威权体制对维护政权稳定与控制力的核心关切。

AI-generated summary

Why It Matters

中国国安部部长陈一新近期发表文章,列举人工智能发展带来的多重国家安全挑战。

Font size

(德国之声中文网)有关人工智能的安全与监管近日来成为国际讨论的话题。而中国国安部部长陈一新新近发表的一篇文章中,列举了人工智能发展可能带来的危险挑战。其中列 in 第一条的就是政治安全。

陈一新写道:“各种敌对势力和别有用心之人滥用深度伪造音视频、AI生成图文、智能水军等生成式人工智能技术,低成本、批量化炮制政治谣言、传播有害信息、煽动对立情绪,对我实施舆论战、认知战,直接威胁我政治安全、制度安全、意识形态安全。”

墨卡托中国研究中心(MERICS)科技与创新项目负责人丽贝卡·阿尔塞萨蒂(Rebecca Arcesati)在接受DW采访时表示,这并非中国国安部首次对AI风险表示关切。“今年7月在上海举行的世界人工智能大会上,中国最高领导人习近平首次公开提及政府或全人类失去对AI系统控制的风险。”

阿尔塞萨蒂表示,应从“一个威权一党制国家执迷于维系权力的角度”来看待这个问题。“这就是为什么中国国安部提及的首要关切就是政治安全。北京担心敌对外国势力会利用 AI 系统在中国内部制造不稳定,例如对公众舆论施加影响,并最终威胁到中国共产党的执政地位。这是他们的首要关切,也解释了为什么中国正日益加大对新兴技术的监管力度,以确保掌控力。”

“微信蠕虫”

陈一新在文中提及美国科技公司近期推出的“神话”(Claude Mythos)、“赛博”(GPT-5.5-Cyber)大模型,表示“部分国家、组织具备快速批量挖掘漏洞、自动串联攻击路径、完成复杂黑客任务的能力,…对我国关键信息基础设施带来严重风险隐患”。

在窃密泄密风险方面,陈一新写道:“境外间谍情报机关深度运用智能网络爬虫、智能数据挖掘、智能画像分析等技术,大肆搜集获取国家关键数据、商业秘密、公民个人隐私等敏感信息数据”。

专家阿尔塞萨蒂提到近日媒体报道的“微信蠕虫”事件。据《纽约时报》9月9日报道,由于先进的人工智能模型,位于加利福尼亚州帕洛阿尔托的安全公司Calif的一个小型研究团队最近仅用一周多时间就构建了一个黑客工具,它能够在微信——中国无处不在的社交媒体和即时通讯平台——上横行肆虐。

阿尔塞萨蒂就此分析说:“微信的底层设计确实保留了专门供中国官方(尤其是网信办 )进行监控的内部接口。但如果一个外国政府或AI系统突然掌握了直接控制中国微信用户账户的能力,那将是完全不同的情景。正如我之前所说,那将对中共的政治安全构成巨大的潜在威胁。”

中国外交部9月15日在回答有关人工智能安全的提问时表示,中国始终坚持发展和安全并重,高度重视人工智能引发的各类内生和衍生风险,致力于筑牢安全底线,不断完善法律法规、政策制度、应用规范、伦理规则,防范滥用恶用,确保人工智能安全、可靠、可控。

Open Questions

  • 中国未来将出台哪些具体的AI安全监管细则?
  • 微信平台将如何应对类似蠕虫的黑客工具威胁?

Related Topics

This article was originally published by 德国之声.

Related Stories

The final of the second "Bay Area Cup" Cybersecurity Competition was held in Guangzhou, with 150 teams competing against each other.
Tech·

The final of the second "Bay Area Cup" Cybersecurity Competition was held in Guangzhou, with 150 teams competing against each other.

As an important event of the 2026 Guangdong Province Cyber ​​Security Publicity Week, the finals of the second "Bay Area Cup" Cyber ​​Security Competition were held on the 14th at the Guangzhou International Campus of South China University of Technology. 450 players from 150 teams launched a 6.5-hour offensive and defensive battle, focusing on actual network security combat and AI security frontiers.

中国新闻网
4 min read
AI-assisted blind glasses and smart guide dogs are accelerating, and regulatory and emotional shortcomings need to be filled.
Developing·

AI-assisted blind glasses and smart guide dogs are accelerating, and regulatory and emotional shortcomings need to be filled.

Visually impaired people in my country face difficulties in daily travel, and technological products such as AI-assisted glasses and smart guide dogs are accelerating the implementation of pilot projects. However, products still face multiple practical challenges and gaps in terms of flexibility, emotional companionship, and legal and regulatory standards.

中国新闻网
11 min read
More on this topic人工智能