insideout Get Pro

08 · The risks of AI systems

Pro

Security Foundations, lesson 8

The risks of AI systems

How AI systems fail or get misused, and the control for each.

Companies now use AI with customers and at work: language models, which write text by predicting what comes next, and models that score things, such as invoices for fraud. They fail in new ways: text that steers them, data that misleads them, answers that are wrong but sure, and voices they can fake. You'll take one company's support assistant and finance team through each, with its control. No AI background needed.

Part of Pro

In this lesson
  1. Text that steers the model
  2. Data that misleads, answers that are wrong
  3. Faces and voices that can be faked

Cybersecurity comes with Pro: every lesson in every module.

Security Foundations overviewSpotted a mistake? Tell us.