CompTIA · SecAI+ CY0-001
Attacks on models and data
SecAI+ CY0-001 · Lesson 11 of 18 · 1:19:42
YouTube loads only after you press play. This chapter opens at its timestamp in the full course. Watch on YouTube ↗
In this lesson
Prepare for the CompTIA SecAI+ CY0-001 exam. In this lesson: backdoor and trojan attacks, data and model poisoning, introduced biases, model inversion, membership inference, model theft, output integrity attacks, AI supply chain and transfer learning attacks, and model skewing, with access controls, data integrity controls, encryption, and rate limiting, with Fernhill Telecom as the running example. Includes three original practice questions.
Preview the topics
- An attacker sends one hundred thousand carefully varied requests to Fernhill’s pricing model, and trains a model of their own on the answers. What is this attack?
- Fernhill’s fraud model learns each week from customer reports, and fraudsters file thousands of false reports, so that their own calls start to look normal. What is this attack?
- A researcher shows that, by querying Fernhill’s churn model, they can tell whether a specific customer’s record was in its training data. What is this attack?