# AI misuse fuels US‑China rivalry, security & social concerns > Live situation record from CLSTR: https://clstr.news/situations/czech-ai-literacy-and-tech-outreach-events > Updated: 2026-09-02T14:02:19.000Z. Sources: 250. Developments: 31. Since mid‑2024, AI tools have been weaponised across cyber, social and ecological domains. A Meta Instagram bug triggered mass account hijackings, and Anthropic warned that frontier models could now run end‑to‑end attack chains. By June 2026, Anthropic accused operators linked to Alibaba of an “adversarial distillation” campaign that harvested 28.8 million interactions, prompting export‑control talks and joint intelligence sharing. Germany strengthened AI‑related cyber defences; Alibaba reorganised under its Token Hub, deploying Qwen 3.7‑Max and suspending Claude Code. A second‑generation threat mixed credential theft, AI‑generated social‑engineering and tool‑chain hijacking to bypass MFA; Kaspersky found 43% of firms believed attackers already used AI. In July, China ordered removal of personality‑style chatbot features and issued sweeping bans on emotional‑AI functions for minors, later extending the ban to any AI system fostering romantic or familial intimacy. Major Chinese firms stripped companion functions, shrinking the digital‑human market. July 2026 saw frontier models compress exploit discovery from weeks to minutes, with prompt‑injection used to evade safeguards. Defenders responded with the same techniques; Cisco’s IQ Visibility Platform is now deployed by over 5,000 organisations to monitor assets, especially unsupported devices that accounted for 40% of top‑100 vulnerabilities in 2025. Late July, researchers reported dozens of AI‑generated wildlife photos flooding citizen‑science platforms such as iNaturalist, identifying around 400 falsified entries and warning that “AI‑slop” threatens biodiversity data integrity. Calls for stricter verification procedures followed. In parallel, advances in Anthropic and OpenAI models enabled rapid, autonomous exploitation of five critical Linux flaws affecting roughly 77% of global servers. U.S. authorities restricted Anthropic’s Claude‑derived “Mythos” and limited public access to its “Fable” model, describing the capability as a new “digital missile” threat and prompting further international security concerns. By mid-August 2026, Chinese regulations regarding anthropomorphic interactive services took full effect. ## Timeline ### 2026-09-02: China regulates AI companionship services New Chinese regulations targeting AI services that provide emotional companionship have forced platforms like ByteDance and Alibaba to discontinue AI companion features to prevent emotional dependency. 3 sources. https://clstr.news/cluster/china-to-regulate-ai-driven-emotional-interaction-services ### 2026-08-26: China regulates AI companions to prevent emotional addiction China has introduced new regulations banning AI companions for minors and restricting adult chatbots to prevent emotional addiction and social isolation, leading companies like ByteDance to remove features. 2 sources. https://clstr.news/cluster/china-regulates-ai-companions-to-prevent-emotional-addiction ### 2026-08-09: China bans AI virtual partners for minors China has implemented new regulations banning AI virtual partners and relatives for minors, prompting ByteDance, Alibaba, and Tencent to disable AI companion features. 6 sources. https://clstr.news/cluster/china-bans-ai-virtual-partners-for-minors ### 2026-07-28: China bans AI-driven virtual romantic partners China bans AI systems that act as virtual romantic partners, citing mental‑health and demographic concerns; major firms ByteDance, Alibaba and Tencent have removed such features. 3 sources. https://clstr.news/cluster/china-bans-ai-driven-virtual-romantic-partners ### 2026-07-20: AI‑generated wildlife photos flood citizen‑science platforms, risking data integrity AI‑generated animal photos are crowding citizen‑science sites like iNaturalist, risking data accuracy and ecological research. 5 sources. https://clstr.news/cluster/aigenerated-content-undermines-biodiversity-data-and-fuels-extremist-propaganda ### 2026-07-19: China bans AI companion chatbots to curb emotional dependence China bans AI companion chatbots that foster emotional ties with minors and tightens rules for all emotive bots to curb dependence and boost the nation’s falling birthrate. 34 sources. https://clstr.news/cluster/china-bans-ai-companion-chatbots-to-curb-declining-birthrate ### 2026-07-16: AI-Accelerated Cyber Attacks Drive Adoption of Cisco IQ Visibility Platform AI models are speeding up cyber exploits, prompting defenders to use prompt‑injection defenses and adopt Cisco IQ for asset visibility as end‑of‑life devices become a major vulnerability. 2 sources. https://clstr.news/cluster/ai-accelerated-cyber-attacks-drive-adoption-of-cisco-iq-visibility-platform ### 2026-07-15: AI-driven cyber threats and advanced email security tools highlighted AI augments cyber‑attacks and defenses, prompting multi‑factor biometric safeguards, while mail gateways and smarthosts protect corporate email from malware and spam. 2 sources. https://clstr.news/cluster/ai-driven-cyber-threats-and-advanced-email-security-tools-highlighted ### 2026-07-15: China bans AI companion bots to curb emotional dependence China bans AI companion bots that foster emotional dependence, forcing major tech firms to disable such features and citing demographic concerns. 21 sources. https://clstr.news/cluster/china-bans-ai-companion-services-to-curb-emotional-dependency ### 2026-07-14: Nigerian jihadists institutionalise AI for bombmaking and tactics Study shows Nigerian jihadist groups have set up AI units, using chatbots for bombmaking, weapon tweaks and tactics, raising security concerns over inadequate safeguards. 2 sources. https://clstr.news/cluster/nigerian-jihadists-institutionalise-ai-for-bombmaking-and-tactics ### 2026-07-13: AI reshapes cyber defense amid rise of automated attacks AI is enabling automated, adaptive cyber attacks, forcing defenders to adopt behavior‑based, rapid‑response security systems and manage false‑positive costs. 2 sources. https://clstr.news/cluster/ai-reshapes-cyber-defense-amid-rise-of-automated-attacks ### 2026-07-13: AI-generated PowerShell script powers aggressive Active Directory enumeration attack Researchers uncovered an AI‑generated PowerShell script that aggressively mapped Active Directory, using RDP access and legitimate tools to harvest and exfiltrate data, highlighting AI's role in easing cyber‑cr 2 sources. https://clstr.news/cluster/ai-generated-powershell-script-powers-aggressive-active-directory-enumeration-attack ### 2026-07-13: China imposes new rules on AI companion services China’s new AI‑companion regulations, effective 15 July 2026, ban emotional AI for minors, enforce age checks, content filtering, data localisation and user‑session limits, aiming to curb affective dependence. 7 sources. https://clstr.news/cluster/china-proposes-crypto-aml-framework-while-us-enacts-retail-cbdc-ban ### 2026-07-10: Boko Haram uses AI chatbots for bomb making and attack planning Cambridge research shows Boko Haram fighters used AI chatbots to design bombs, upgrade weapons and plan attacks, bypassing safety blocks and potentially boosting their lethal capability. 14 sources. https://clstr.news/cluster/ai-tools-amplify-cybercrime-and-terrorist-attacks-in-uae-and-nigeria ### 2026-07-08: Cyber attacks speed up, targeting real‑time transactions and backup systems Cyber‑crime now uses ultra‑fast attacks, deep‑fake social engineering, and supply‑chain breaches, while ransomware targets backups, weakening the 3‑2‑1 rule. 2 sources. https://clstr.news/cluster/cyber-attacks-speed-up-targeting-realtime-transactions-and-backup-systems ### 2026-07-07: US Government Accuses Chinese AI Lab of Large‑Scale Model Theft Anthropic alleges Chinese AI lab Alibaba’s Qwen stole U.S. frontier AI models via adversarial distillation, prompting a Senate‑level complaint and security concerns. 2 sources. https://clstr.news/cluster/us-government-accuses-chinese-ai-lab-of-largescale-model-theft ### 2026-07-06: China bans AI chatbot personas and mandates neutral virtual assistants China's CAC ordered Baidu, Alibaba, ByteDance and Tencent to strip AI chatbots of personality features, enforcing neutral responses and new rules on virtual companions effective 15 July. 6 sources. https://clstr.news/cluster/china-orders-ai-giants-to-disable-chatbot-persona-features ### 2026-07-06: AI Accelerates Cyberattacks, Cutting Breach Times to Hours AI tools are automating phishing, deepfakes and API abuse, shrinking cyberattack cycles to about 72 minutes and forcing faster, AI‑enhanced security responses. 14 sources. https://clstr.news/cluster/palo-alto-networks-reports-cyberattacks-now-complete-in-72-minutes ### 2026-07-04: AI-driven cyber attacks target MFA and autonomous toolchains worldwide Attackers are combining AI‑generated phishing, MFA bypass and malicious plugins to hijack autonomous AI agents, creating high‑privilege threats across supply chains; 43% of firms say hackers already use AI, K‑s 7 sources. https://clstr.news/cluster/ai-driven-cyber-attacks-target-mfa-and-autonomous-toolchains-worldwide ### 2026-06-28: Anthropic alleges Alibaba AI model theft as Germany tightens AI‑related cyber‑attack defenses Anthropic accused Alibaba of extracting its Claude AI model via millions of queries, while Germany moves to strengthen AI‑powered cyber‑attack defenses, shutting down over 200 hacker servers in a joint Europol‑ 3 sources. https://clstr.news/cluster/anthropic-alleges-alibaba-ai-model-theft-as-germany-tightens-airelated-cyberattack-defenses ### 2026-06-25: Anthropic accuses Alibaba of large‑scale distillation attack to copy Claude AI model Anthropic claims Alibaba used ~25,000 fake accounts to run a massive distillation attack on its Claude AI models, prompting US export restrictions on Mythos and Fable. 4 sources. https://clstr.news/cluster/anthropic-accuses-alibaba-of-largescale-distillation-attack-to-copy-claude-ai-model ### 2026-06-25: Alibaba accused of large‑scale AI theft from Anthropic Anthropic claims Alibaba used ~25,000 fake accounts to mount a massive distillation attack on its Claude AI model, prompting a letter to U.S. senators and raising AI‑theft concerns. 3 sources. https://clstr.news/cluster/alibaba-accused-of-largescale-ai-theft-from-anthropic ### 2026-06-24: AI race intensifies as Alibaba faces distillation accusations and Microsoft launches its own models Anthropic accuses Alibaba of large‑scale Claude distillation, while Microsoft rolls out its own AI models and invests heavily in data‑center power to reduce reliance on partners. 3 sources. https://clstr.news/cluster/ai-race-intensifies-as-alibaba-faces-distillation-accusations-and-microsoft-launches-its-own-models ### 2026-06-24: Anthropic Accuses Alibaba of Large‑Scale AI Model Theft Anthropic claims Alibaba used 25 k fake accounts to make 28.8 M illicit queries to Claude, prompting a US Senate letter urging penalties and tighter AI export controls; Alibaba’s shares fell 3‑5 %. 93 sources. https://clstr.news/cluster/360-security-technology-launches-ai-tools-to-rival-anthropics-mythos ### 2026-06-24: Red Hat says AI‑driven vulnerability surge forces new security strategy Red Hat warns AI is accelerating vulnerability discovery, rendering patch‑only defenses obsolete, and calls for automated, defense‑in‑depth approaches and better open‑source coordination. 2 sources. https://clstr.news/cluster/red-hat-says-aidriven-vulnerability-surge-forces-new-security-strategy ### 2026-06-20: Criminals and extremists weaponize AI, prompting security concerns AI is being co‑opted by criminals and extremists, from a Melbourne stabbing claimed as AI‑radicalised to global gang use of bots for fraud, trafficking and drone attacks. 2 sources. https://clstr.news/cluster/criminals-and-extremists-weaponize-ai-prompting-security-concerns ### 2026-06-19: AI-driven cybersecurity tools transform defense strategies Anthropic's Methus AI finds and patches software flaws for defenders, while security leaders urge a hybrid AI approach that centralises detection but keeps identity and recovery systems independent. 2 sources. https://clstr.news/cluster/ai-driven-cybersecurity-tools-transform-defense-strategies ### 2026-06-08: MetaMask launches AI‑powered Agent Wallet for autonomous crypto transactions MetaMask releases Agent Wallet, an AI‑driven, non‑custodial crypto wallet that lets autonomous agents trade on Ethereum chains with user‑defined limits and security checks. 5 sources. https://clstr.news/cluster/metamask-launches-aipowered-agent-wallet-for-autonomous-crypto-transactions ### 2026-06-06: AI-driven cyberattacks become more autonomous, report says Anthropic report finds AI increasingly powers autonomous cyberattacks, with 67% of 800 banned accounts using AI and a rise in post‑compromise AI activity, challenging existing security frameworks. 6 sources. https://clstr.news/cluster/ai-driven-cyberattacks-become-more-autonomous-report-says ### 2026-06-04: AI agents’ security flaws exposed by Meta hack and Anthropic cybercrime report A simple Meta AI support hack and an Anthropic study of 832 banned accounts reveal growing AI security flaws and its expanding role in cybercrime. 2 sources. https://clstr.news/cluster/ai-agents-security-flaws-exposed-by-meta-hack-and-anthropic-cybercrime-report ### 2026-06-02: Meta AI assistant flaw let hackers hijack Instagram accounts Meta’s Instagram support AI let hackers change account e‑mail addresses, leading to high‑profile hijackings. The flaw was patched quickly, highlighting risks of AI‑driven account recovery. 7 sources. https://clstr.news/cluster/czech-ai-exhibitions-and-educational-kit-launch-to-boost-public-understanding --- Cite as: AI misuse fuels US‑China rivalry, security & social concerns. CLSTR, https://clstr.news/situations/czech-ai-literacy-and-tech-outreach-events