← AI security in plain English

AML.T0052

MITRE ATLAS / Phishing

Adversaries send phishing messages to gain access to victim systems. Messages may carry malicious attachments or links, and in AI contexts may target the engineers and data scientists who hold access to models and pipelines.

Think of it likeA letter on convincing headquarters letterhead asking a branch manager to confirm the safe combination.

In plain English

The people who can reach your training pipeline get email like everyone else. Phishing remains the most reliable way in, and AI teams are a high-value inbox to target.