[CVE-2025-32711] AI Security Risks and Vulnerabilities Enterprises Can’t Ignore

CVE

CVE-2025-32711

Published

2025-09-30 18:56

Collected

2026-07-02 04:27

---
id: 1174
title: "[CVE-2025-32711] AI Security Risks and Vulnerabilities Enterprises Can’t Ignore"
cve: CVE-2025-32711
source_domain: hackerone.com
source_url: "https://www.hackerone.com/blog/ai-security-risks-and-vulnerabilities"
published_at: "2025-09-30 18:56"
tags: ["ai-security", "supply-chain", "agentic-ai", "mcp", "prompt-injection", "indirect-prompt-injection", "enterprise-security", "data-exfiltration", "echoleak", "cve-2025-32711", "microsoft-365-copilot"]
phying_url: "https://phying.com/item/1174"
lang: en
---

# [CVE-2025-32711] AI Security Risks and Vulnerabilities Enterprises Can’t Ignore

## Summary

Enterprises adopting AI face eroding trust boundaries: indirect prompt injection like EchoLeak (CVE-2025-32711) can hijack Microsoft 365 Copilot, while MCP and agentic AI expand the attack surface.

## Intelligence Metadata

- **Sources:** hackerone.com
- **CVE:** CVE-2025-32711
- **Published:** 2025-09-30 18:56
- **Original:** https://www.hackerone.com/blog/ai-security-risks-and-vulnerabilities
- **Phying News:** https://phying.com/item/1174