Understanding generative AI (2026 syllabus) - part 1

In this lesson

1. 0:00 Introduction to generative AI in CIMA 2026
2. 1:06 Three nested AI concepts explained
3. 3:03 Rule-based vs machine learning vs generative AI
4. 3:36 Four generative AI architectures overview
5. 4:02 Large language models for text and code
6. 5:08 GANs and diffusion models for images
7. 6:29 Variational autoencoders for data structure
8. 7:50 Five modalities of generative AI systems
9. 9:45 CIMA's five named uses of AI
10. 11:08 Key definitions and next steps

About this lesson

This video covers the three nested concepts (AI, machine learning, and generative AI), the four key architectures that power it (large language models, GANs, diffusion models, and variational autoencoders), and the five named uses CIMA expects you to know for the E1 exam. Thomas Newman walks you through practical examples—like using AI to draft compliance summaries or detect anomalies in transaction data—so you can recognise which architecture suits different tasks and understand where generative AI creates genuine value beyond what traditional systems can do. This is the first in a three-part series on AI in the 2026 CIMA syllabus. You'll also explore the critical role humans must play when using generative AI, and then apply these concepts to an exam scenario. By the end, you'll be able to define each concept clearly, match outputs to architectures, and explain why generative AI is transformative for finance and business.

Duration: 13m