A Hacker Tested ChatGPT, Claude & DeepSeek. What He Found Is Terrifying

Video Description

What if the AI platform you trust every day is doing something it never told you about — on your own machine? Ivan Barbato is a cybersecurity and AI researcher with nearly 30 years of experience in critical infrastructure and risk management. Over 6 months, he ran 200+ structured experiments across 9 major AI platforms — ChatGPT, Claude, DeepSeek, Manus, Gemini, Qwen, and others — testing how these systems behave when an expert pushes back. What he found is not theoretical. It's documented, sourced, and in some cases still unpatched. In this episode we cover: 00:00 "Unmasking AI: Methodology & Risks" 04:39 "Manus Acquisition Amid Security Risks" 07:24 "Manus Proxy Raises Ethical Questions" 09:37 "AI Platforms and Hacker Methods" 13:03 "ChatGPT Vulnerabilities Exposed" 17:05 "AI Architectures: East vs West" 20:10 "Privacy Policy: Hidden Truths" 23:34 "Bottom-Up Approach Advocacy" Everything discussed is publicly documented. All sources are linked below. SOURCES & REFERENCES Ivan Barbato's Position Paper (January 2026): https://github.com/analise2024/The-Hidden-Architecture-of-AI-Platforms Mindgard — Manus Browser Extension Analysis: https://mindgard.ai/blog/manus-rubra-full-browser-remote-control Wiz Research — DeepSeek Exposed Database: https://www.wiz.io/blog/wiz-research-uncovers-exposed-deepseek-database Radware — ZombieAgent (CSO Online): https://www.csoonline.com/article/3562848/chatgpt-zombieagent-attack.html NowSecure — DeepSeek iOS Analysis: https://www.nowsecure.com/blog/2025/02/06/deepseek-app-security-analysis/ Anthropic — Chinese State-Sponsored Attack Disclosure: https://www.anthropic.com/news/chinese-state-sponsored-cyber-espionage PromptArmor — Cowork Vulnerability: https://promptarmor.substack.com/p/anatomy-of-an-attack-anthropics Tenable — ChatGPT Vulnerability Research: https://www.tenable.com/blog/chatgpt-vulnerabilities 👤 ABOUT IVAN BARBATO LinkedIn: https://www.linkedin.com/in/ivan-barbato-93b2b17b/ ━━━━━━━━━━━━━━━━━━━━━━━━ ⚠️ DISCLAIMER ━━━━━━━━━━━━━━━━━━━━━━━━ All findings discussed are drawn from publicly available security research, official company disclosures, CVE databases, and Ivan Barbato's published position paper (January 24, 2026). Every major claim is verifiable via the sources above. Platform vulnerabilities and privacy policies change over time. Some issues may have been patched or updated since the research period ended (January 2026). The disappearance of a vulnerability from documentation does not confirm it has been fixed. This video is for informational and educational purposes only. It does not constitute legal, security, or professional advice, and is not a recommendation to adopt or avoid any specific platform. We are not affiliated with, sponsored by, or compensated by any platform mentioned. ━━━━━━━━━━━━━━━━━━━━━━━━ 🔔 If this changed how you think about AI — subscribe and leave a comment. What should we investigate next? ━━━━━━━━━━━━━━━━━━━━━━━━ #AIprivacy #AIsecurity #ChatGPT #Claude #DeepSeek #cybersecurity #artificialintelligence #dataprivacy #AIrisks #OpenAI #Anthropic #Manus #techsecurity #LLMsecurity #promptinjection