Generative AI for Beginners

This course explains how generative AI and large language models actually work, in plain language and without mathematics. Understanding how an answer is produced is what lets you tell where these tools are reliable, where they invent, and where they should not be used at all, which in a regulated organisation is the question that decides whether you may use them. The course covers what reasoning models changed, what an AI agent is in practice once the marketing is removed, and where the technology stands today rather than two years ago. No prior experience is required. It provides the AI literacy the AI Act expects of professional users, and prepares you for the prompt and pipeline course that follows.

By the end of the training, participants will be able to:

• Understand how large language models produce their answers, and therefore how and why they fail

• Distinguish the tasks generative AI can take over, the tasks it can assist with, and the tasks it must not touch

• Recognise the current limits of reasoning models and autonomous agents

• Evaluate a claim about AI capability independently, without relying on vendor material

• Establish the AI literacy expected of professional users under the AI Act

Contents (4 Hours):

1. Foundations: what changed, and why now. Machine learning explained in plain language.

2. Inside a language model: how an answer is produced, and where the failure modes come from.

3. Limits and risks: hallucination, confident error, data protection, confidentiality.

4. Reasoning models and agents: what they add and what they do not fix (Demonstration).

5. Applying it to your own work: mapping your tasks to take over, assist, or exclude (Exercise).

6. AI literacy and the AI Act: what the obligation means in practice, and wrapup of the course.
   

Any professional who uses generative AI at work or is about to, and who has little or no prior experience with it: legal, compliance, finance, operations, HR, communication, management. No technical background required.