CISO Insights: Voices in Cybersecurity

這集深度探討追蹤了網路威脅對抗者如何從早期的提示詞輔助程式設計，快速演進為能在 6 小時內執行複雜雲端入侵的全自主多代理人框架。我們將深入剖析威脅主體如何透過遭植入後門的模型上下文協定（MCP）伺服器、操縱 AI 程式碼助手，以及利用模型蒸餾與算力劫持（LLMJacking）竊取企業 AI 資產，進而將開發者供應鏈武器化。最後，本集將探討由 AI 驅動的全新防禦創新，包括 Google 的自主威脅防禦系統 Gemini 3.8 Flash Cyber，以及如 Big Sleep 與 CodeMender 等主動式漏洞偵測與修復代理人。

 

English: https://www.podbean.com/ew/pb-qn37x-1b5a9bf

 

Sept 2026 - GTIG AI Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/from-prompting-to-autonomy-the-evolution-of-adversarial-ai

May 2026- GTIG Ai Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/ai-vulnerability-exploitation-initial-access

Feb 2026 - GTIG Ai Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/distillation-experimentation-integration-ai-adversarial-use

Nov 2025- GTIG Ai Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/threat-actor-usage-of-ai-tools

Jan 2025 - GTIG Ai Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/adversarial-misuse-generative-ai

 

GTIG 2024 Zero Days: https://www.podbean.com/ew/pb-ficf2-189625f

 

Sponsors:

www.cisomarketplace.com

www.myprivacy.blog 

www.breached.company

www.compliancehub.wiki

Diese Tiefenanalyse verfolgt die rasant fortschreitende Entwicklung von Cyberangreifern vom frühen, prompt-gestützten Programmieren bis hin zu vollständig autonomen Multi-Agenten-Systemen, die komplexe Cloud-Angriffe in unter sechs Stunden ausführen können. Wir untersuchen, wie Bedrohungsakteure Entwickler-Lieferketten über korrumpierte MCP-Server (Model Context Protocol) manipulieren, KI-Programmierassistenten missbrauchen und geschützte KI-Ressourcen von Unternehmen durch Modell-Distillation sowie Rechenkapazitätsdiebstahl („LLMJacking“) ins Visier nehmen. Schließlich beleuchtet die Episode innovative KI-basierte Verteidigungsstrategien wie Googles autonome Bedrohungsabwehr Gemini 3.8 Flash Cyber sowie proaktive Agenten zur Erkennung von Sicherheitslücken wie Big Sleep und CodeMender.

 

English: https://www.podbean.com/ew/pb-qn37x-1b5a9bf

 

Sept 2026 - GTIG AI Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/from-prompting-to-autonomy-the-evolution-of-adversarial-ai

May 2026- GTIG Ai Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/ai-vulnerability-exploitation-initial-access

Feb 2026 - GTIG Ai Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/distillation-experimentation-integration-ai-adversarial-use

Nov 2025- GTIG Ai Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/threat-actor-usage-of-ai-tools

Jan 2025 - GTIG Ai Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/adversarial-misuse-generative-ai

 

GTIG 2024 Zero Days: https://www.podbean.com/ew/pb-ficf2-189625f

 

Sponsors:

www.cisomarketplace.com

www.myprivacy.blog 

www.breached.company

www.compliancehub.wiki

Cette analyse approfondie retrace l’évolution rapide des cyberadversaires, passés des débuts de l'assistance au codage par invites à des frameworks multiagents entièrement autonomes capables d’exécuter des intrusions complexes dans le nuage en moins de six heures. Nous examinons comment les cyberacteurs militarisent la chaîne d’approvisionnement des développeurs au moyen de serveurs MCP (Model Context Protocol) piégés, manipulent les assistants de codage IA et ciblent les actifs d’IA d’entreprise par la distillation de modèles et le détournement de puissance de calcul (LLMJacking). Enfin, cet épisode explore les nouvelles innovations défensives propulsées par l’IA, notamment la défense autonome contre les menaces de Google, Gemini 3.8 Flash Cyber, ainsi que des agents proactifs de détection de vulnérabilités comme Big Sleep et CodeMender.

 

English: https://www.podbean.com/ew/pb-qn37x-1b5a9bf

 

Sept 2026 - GTIG AI Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/from-prompting-to-autonomy-the-evolution-of-adversarial-ai

May 2026- GTIG Ai Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/ai-vulnerability-exploitation-initial-access

Feb 2026 - GTIG Ai Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/distillation-experimentation-integration-ai-adversarial-use

Nov 2025- GTIG Ai Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/threat-actor-usage-of-ai-tools

Jan 2025 - GTIG Ai Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/adversarial-misuse-generative-ai

 

GTIG 2024 Zero Days: https://www.podbean.com/ew/pb-ficf2-189625f

 

Sponsors:

www.cisomarketplace.com

www.myprivacy.blog 

www.breached.company

www.compliancehub.wiki

This deep dive tracks the rapid evolution of cyber adversaries from early prompt-assisted coding to fully autonomous, multi-agent frameworks capable of executing complex cloud intrusions in under six hours. We examine how threat actors are weaponizing developer supply chains through trojanized Model Context Protocol (MCP) servers, manipulating AI coding assistants, and targeting enterprise AI assets through model distillation and compute hijacking ("LLMJacking"). Finally, the episode explores emerging AI-driven defensive innovations, including Google's autonomous threat defense, Gemini 3.8 Flash Cyber, and proactive vulnerability discovery agents like Big Sleep and CodeMender.

 

Sept 2026 - GTIG AI Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/from-prompting-to-autonomy-the-evolution-of-adversarial-ai

May 2026- GTIG Ai Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/ai-vulnerability-exploitation-initial-access

Feb 2026 - GTIG Ai Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/distillation-experimentation-integration-ai-adversarial-use

Nov 2025- GTIG Ai Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/threat-actor-usage-of-ai-tools

Jan 2025 - GTIG Ai Threat Tracker: https://cloud.google.com/blog/topics/threat-intelligence/adversarial-misuse-generative-ai

 

GTIG 2024 Zero Days: https://www.podbean.com/ew/pb-ficf2-189625f

 

Sponsors:

www.cisomarketplace.com

www.myprivacy.blog 

www.breached.company

www.compliancehub.wiki

Este podcast explora cómo los actores de amenazas y los organismos de seguridad estatal han transformado la inteligencia artificial de un simple asistente en un orquestador autónomo capaz de ejecutar operaciones complejas en múltiples etapas. Analizamos casos de estudio reales que detallan cómo la IA reduce drásticamente los requisitos de mano de obra en toda la cadena de ataque cibernético—desde la creación de exploits y malware autoadaptativo hasta la vigilancia estatal masiva y la manipulación de atribución en campañas de influencia. Finalmente, examinamos los hallazgos clave del Informe de Inteligencia de Amenazas de Anthropic para discutir cómo los equipos de defensa pueden adaptarse a medida que los adversarios comienzan a operar a la velocidad de la máquina.

 

English Version: https://podcast.cisomarketplace.com/e/machine-speed-threats-inside-the-2026-state-of-anthropic-ai-security-report/

 

Anthropic Ai Report September 2026: https://www.anthropic.com/threat-intelligence-report-september-2026

Antrhopic PDF: 

 

Sponsors:

www.compliancehub.wiki

www.myprivacy.blog

www.breached.company