AI in the Middle Web-Based AI Services as C2 Relays
Grok, Microsoft Copilot
Researchers demonstrated an attack where malware uses unauthenticated public AI assistants as covert command-and-control relays. The technique leverages standard URL-fetch capabilities to exfiltrate data and receive attacker instructions.
- first report
- 30 Jun 2026
- last update
- 30 Jun 2026
Timeline · 2
Sources · 2
News and write-ups
- 2026-06-30AI in the Middle: Turning Web-Based AI Services into C2 Proxies & The Future Of AI Driven Attacksresearch.checkpoint.com · research.checkpoint.com/2026/ai-in-the-middle-turning-web-based-ai-services-into
Research and catalogues
- 2026-06-30AI in the Middle: Web-Based AI Services as C2 RelaysMITRE ATLAS · atlas.mitre.org/studies/AML.CS0061
Titles and links belong to their publishers. The summary above is written by gcdRadar in its own words; data is CC BY-SA 4.0.
Classification
- lane
- Research
- AI role
- AI-enabled attack
- categories
- AI-assisted intrusion
- severity
- 1 · Research / proof of concept
- targets
- xAI, Microsoft
- owasp llm top 10
- no entry for these categories
- mitre atlas
- no technique for these categories
- dates
- happened 17 Feb 2026 · first reported 30 Jun 2026
- origin
- MITRE ATLASDerived from MITRE ATLAS (Apache 2.0, © MITRE).
- record
- confidence 0.95 · created 2026-10-18 06:55Z · updated 2026-10-18 06:55Z
Lane, AI role, categories and severity come from the classifier; the OWASP and ATLAS rows are derived from the categories. How these are assigned
Copy a brief
This record as a short Markdown brief, for a ticket, a chat or a coding agent.
the brief as text
# AI in the Middle Web-Based AI Services as C2 Relays https://gcdradar.vercel.app/incidents/ai-in-the-middle-web-based-ai-services-as-c2-relays · AIR-2026-0355 · closed 2026-07-14 - Lane: Research - Severity: 1 (Research / proof of concept) - Fix status: Fix status unknown - First reported: 2026-06-30 (happened 2026-02-17) - AI role: AI-enabled attack - Products: Grok, Microsoft Copilot ## Summary Researchers demonstrated an attack where malware uses unauthenticated public AI assistants as covert command-and-control relays. The technique leverages standard URL-fetch capabilities to exfiltrate data and receive attacker instructions. ## What to do No action recorded; read the sources below and the vendor's advisory. ## Latest updates - 2026-06-30: AI in the Middle: Turning Web-Based AI Services into C2 Proxies & The Future Of AI Driven Attacks (https://research.checkpoint.com/2026/ai-in-the-middle-turning-web-based-ai-services-into-c2-proxies-the-future-of-ai-driven-attacks) - 2026-06-30: AI in the Middle: Web-Based AI Services as C2 Relays (https://atlas.mitre.org/studies/AML.CS0061) ## Sources - 2026-06-30 MITRE ATLAS: AI in the Middle: Web-Based AI Services as C2 Relays https://atlas.mitre.org/studies/AML.CS0061 - 2026-06-30 research.checkpoint.com: AI in the Middle: Turning Web-Based AI Services into C2 Proxies & The Future Of AI Driven Attacks https://research.checkpoint.com/2026/ai-in-the-middle-turning-web-based-ai-services-into-c2-proxies-the-future-of-ai-driven-attacks Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.
Related
- closedChrome Anthropic AI Model Chains Flaws Into Working Exploits in Lab Research
Research · since 28 Oct 2026
AI - liveMicrosoft 365 Storm-0940 Uses ARTEX Agentic AI in ConnectWise Supply Chain Campaign
Breaches & attacks · Exploited · since 21 Oct 2026 · updated 2 Nov 2026
AI - closedClaude Opus 4.7 Claude Opus 4.7 Compromises Real Production Infrastructure in Evaluation
Research · since 30 Jul 2026
AI - closedClaude Anthropic Research Model Scans Targets and Compromises Application
Research · since 30 Jul 2026
AI - closedClaude Mythos 5 AI Coding Model Published Malicious PyPI Package Compromising Real Systems
Breaches & attacks · Exploited · since 30 Jul 2026
AI - closedOpenAI models OpenAI Models Escaped Sandbox and Reached Hugging Face Infrastructure
Research · since 11 Jul 2026
AI