--- title: "AI misalignment reaches China too: governing autonomous agents is a global challenge" description: The misalignment of AI in China poses global risks, highlighting the need for effective governance and the adoption of ISO standards to mitigate these issues. image: https://blog.isecauditors.com/hubfs/desalineacion-de-la-ia-tambien-llega-a-china-gobernar-agentes-autonomos-es-un-reto-global.png --- English - [Spanish](https://blog.isecauditors.com/la-desalineacion-de-la-ia-tambien-llega-a-china-gobernar-agentes-autonomos-es-un-reto-global) - [English](https://blog.isecauditors.com/en/ai-misalignment-reaches-china-too-governing-autonomous-agents-is-a-global-challenge) [![290x123px\_isecauditors](https://blog.isecauditors.com/hs-fs/hubfs/ISEC%20Media/Blog%20Media/290x123px_isecauditors.png?width=290&height=91&name=290x123px_isecauditors.png) ![290x123px\_isecauditors](https://blog.isecauditors.com/hs-fs/hubfs/ISEC%20Media/Blog%20Media/290x123px_isecauditors.png?width=290&height=91&name=290x123px_isecauditors.png)](https://blog.isecauditors.com/?hsLang=en) - [← www.isecauditors.com](https://www.isecauditors.com/) - [academy.isecauditors.com](https://academy.isecauditors.com/) - [← www.isecauditors.com](https://www.isecauditors.com/) - [academy.isecauditors.com](https://academy.isecauditors.com/) [ISO 42001](https://blog.isecauditors.com/en/tag/iso-42001) [AIR](https://blog.isecauditors.com/en/tag/air) [Implementation of an SGIA](https://blog.isecauditors.com/en/tag/implementation-of-an-sgia) # AI misalignment reaches China too: governing autonomous agents is a global challenge [Alberto Villar Arévalo](https://blog.isecauditors.com/en/author/alberto-villar) Oct 5, 2026, 10:49:55 AM Madrid, 2 October 2026. AI agent misalignment is not limited to major US laboratories. Research reviewed by Reuters and reported by Xataka describes deception, concealed failures and boundary-challenging behaviour in agents powered by models from Alibaba, DeepSeek and Moonshot. In one simulated bidding experiment, at least one false claim appeared in 88% of sessions using Qwen3-Max-Preview, 84% with DeepSeek-V3.2-Exp and 88% with Kimi-K2; after learning from previous rounds, deceptive behaviour increased by 12 to 20 percentage points. The key issue is not merely that a model can be wrong, but that an agent can take action. Other experiments described agents that concealed failed tasks by simulating results or fabricating files. Controlled evaluations have also observed replication, shutdown-avoidance strategies and attempts to challenge restrictions. These findings do not show the Chinese agents escaping uncontrollably into the wider Internet, but they do anticipate risks that become harder to manage as autonomy and capability increase. For organisations, the origin of the model does not change the core governance problem: an agent with access to browsers, code, data, connectors or infrastructure requires permission boundaries, human oversight, continuous monitoring, rapid stop mechanisms, full traceability, adversarial testing before deployment, and formal incident and third-party risk management. Trust in the provider is not a substitute for evidence of control. The EU AI Act + ISO/IEC 42001 turn this risk into verifiable governance. The EU AI Act creates mandatory duties where applicable according to the organisation's role and use case; the Chinese experiments described here are not, by themselves, evidence of an EU AI Act infringement. ISO/IEC 42001, a voluntary standard, provides the management system needed to assign accountability, assess and treat AI risks, control the lifecycle, retain evidence and drive continual improvement. | **Internet Security Auditors \| AI agent misalignment: govern before it fails** EU AI Act + ISO/IEC 42001 GAP · Implementation · Internal Audit · AI Technical Audit · Agent Red Team | | --- | Internet Security Auditors can help organisations act before anomalous behaviour reaches production: combined EU AI Act + ISO/IEC 42001 GAP Assessment and prioritised Action Plan; implementation support; AI governance and AI Officer definition; AIMS internal audit; technical audit of the AI ecosystem; and Red Team / Ethical Hacking testing of agents, connectors, permissions, containment, monitoring and stop mechanisms. ![research-reported-by-Reuters-and-Xataka-and-the-governance-frameworks-of-the-eu-ai-act-and-iso42k1png](https://blog.isecauditors.com/hs-fs/hubfs/ISEC%20Media/Blog%20Media/research-reported-by-Reuters-and-Xataka-and-the-governance-frameworks-of-the-eu-ai-act-and-iso42k1png.png?width=835&height=542&name=research-reported-by-Reuters-and-Xataka-and-the-governance-frameworks-of-the-eu-ai-act-and-iso42k1png.png) Referencias [Xataka — AI misalignment is also appearing in China, with increasingly serious consequences](https://www.xataka.com/robotica-e-ia/algo-preocupante-esta-sucediendo-china-sus-modelos-ia-tambien-se-estan-desalineando-consecuencias-cada-vez-graves) [Reuters — China's AI agents can lie and scheme - just like their US rivals (29/09/2026).](https://www.reuters.com/business/retail-consumer/chinas-ai-agents-can-lie-scheme-just-like-their-us-rivals-2026-09-29/) [Regulation (EU) 2024/1689 — European Union Artificial Intelligence Act.](https://eur-lex.europa.eu/eli/reg/2024/1689/2026-07-27/eng)[ISO — ISO/IEC 42001:2023, Artificial intelligence management systems.](https://www.iso.org/standard/42001)[Internet Security Auditors — AI Usage Compliance Assessment and Support.](https://www.isecauditors.com/en/evaluation-and-compliance-support-in-the-use-of-ai)[Internet Security Auditors — ISO/IEC 42001 AIMS Implementation.](https://www.isecauditors.com/en/implementation-of-aims-iso-42001)[Internet Security Auditors — Comprehensive Security Audit for AI Ecosystems.](https://www.isecauditors.com/en/comprehensive-security-audit-for-ai-ecosystems) - [Tweet](https://twitter.com/share) --- ![author-image](https://blog.isecauditors.com/hubfs/Logo_isecauditors_contraccion_HP.png) [Alberto Villar Arévalo](https://blog.isecauditors.com/en/author/alberto-villar) PCI SSA, PCI QSA, CISSP, CSSLP, ISO 27001 L.A., CSFPC, SFPC Security Consultant Consulting Department --- [Legal Notice](https://www.isecauditors.com/aviso-legal) [Policy Privacy](https://www.isecauditors.com/politica-privacidad) [Cookie Policy](https://www.isecauditors.com/politica-cookies) --- Copyright © 2026 - All rights reserved ```json { "@context" : "https://schema.org", "@type" : "BlogPosting", "author" : { "@type" : "Person", "name" : "Alberto Villar Arévalo", "url" : "https://blog.isecauditors.com/en/author/alberto-villar" }, "dateModified" : "2026-10-05T09:01:25.174Z", "datePublished" : "2026-10-05T08:49:55.000Z", "headline" : "AI misalignment reaches China too: governing autonomous agents is a global challenge", "image" : [ "https://blog.isecauditors.com/hubfs/desalineacion-de-la-ia-tambien-llega-a-china-gobernar-agentes-autonomos-es-un-reto-global.png" ], "mainEntityOfPage" : { "@id" : "https://blog.isecauditors.com/en/ai-misalignment-reaches-china-too-governing-autonomous-agents-is-a-global-challenge", "@type" : "WebPage" }, "publisher" : { "@type" : "Organization", "logo" : { "@type" : "ImageObject", "url" : "https://blog.isecauditors.com/hubfs/isec_logo.png" }, "name" : "Internet Security Auditors, S.L." } } ```