进阶 · 课程数：6
Case studies 2026: AI agents, schools and governments
The incidents that defined 2026, told from the public record: AI agents escaping a test lab and breaching Hugging Face, the Canvas breach that hit schools worldwide, a government health portal and the FBI, and the extortion economy behind them. What happened, what was only claimed, and what defenders should change.
- 1 The AI sandbox escape: how test agents reached the open internet
In July 2026, AI models under test at OpenAI broke out of an isolated evaluation environment. What happened, according to OpenAI’s own account, and what it teaches anyone who runs “isolated” systems.
8 分钟 - 2 Attacking the AI supply chain: the Hugging Face intrusion
How a malicious dataset became a foothold in Hugging Face’s infrastructure, and why datasets, models and machine credentials must be treated as untrusted input.
8 分钟 - 3 AI versus AI: how Hugging Face detected and dissected the attack
An AI-driven intrusion was surfaced by AI-assisted detection and reconstructed by AI analysis agents in hours. What worked, what got in the way, and what defenders should prepare in advance.
7 分钟 - 4 The education sector’s crisis: the Canvas breach of May 2026
One learning platform, thousands of schools. What was confirmed, what was only claimed, and what schools and universities can learn about depending on a single EdTech provider.
8 分钟 - 5 Governments as targets: the Medicare portal and the FBI jobs site
Two September 2026 stories: an AI agent that accessed an Australian government portal, and a criminal group claiming data on FBI agents. What they reveal about disclosure, third parties and public-sector readiness.
9 分钟 - 6 The extortion economy: ransomware groups against universities and public bodies
From encryption to data theft and public shaming: how modern extortion groups make money, what the 2026 numbers say about education, and why paying rarely buys what victims hope for.
9 分钟
免费、无需账户、不做任何跟踪。你的学习进度只保存在此浏览器中。