AI

Introduction to Generative AI - Exploring the Basics

Introduction to Generative AI - Exploring the Basics
Quick answer

Generative AI is software that learns patterns from huge amounts of data and then creates new text, images or speech from a prompt. The engine behind most chat tools is the large language model (LLM). It can summarise, classify, translate and write, but it can also be wrong or biased, so you check what it gives you. This guide follows the eight units of Microsoft's free "Introduction to generative AI - Exploring the basics" module (about 22 minutes, beginner level) and explains each one in plain English.

Microsoft Learn offers a short, free beginner module called Introduction to generative AI - Exploring the basics. It is aimed at business users and educators, and it uses Microsoft Copilot for its examples. We have walked through all eight units below, in the same order and under the same headings, and added a few notes from our own side. The explanations are our own words. For the original lessons, videos and the graded assessment, use the module on Microsoft Learn.

Introduction

The module opens with a simple promise: if you understand how generative AI works, you can use it more responsibly. That means judging its answers instead of accepting them, spotting where it is likely to go wrong, and picking the right tool for the job.

By the end, the module says you should be able to:

  • explain what generative AI is,
  • describe what large language models do and why they matter,
  • assess your own critical thinking when working with AI, and
  • compare different kinds of generative AI models and know when to use each.

What is generative AI?

Generative AI is a type of artificial intelligence that studies patterns in large, complex sets of data and uses them to produce something new: text, speech or images. Older, "predictive" AI mostly forecasts from existing data, such as flagging a likely spam email. Generative AI goes a step further and creates original content. Because it is generating, not looking up, you will usually get a slightly different answer each time you ask the same question.

The module names two main kinds:

  • Natural language generation. Models that write text the way people do. GPT (Generative Pretrained Transformer) is the best-known example. You give it a prompt and it replies in natural language.
  • Text to image. Models that turn a written description into a picture, such as icons, patterns, illustrations, artwork, imagined photos or 3D-style objects.

What are large language models (LLMs)?

A large language model is a neural network trained on a very large collection of text, including books, articles and websites. Because of that training it can hold a conversation, answer questions, write code, and work with text in many other ways. GPT is one type of LLM.

The module groups what LLMs do into four everyday jobs:

JobWhat it meansExample
ClassificationSorting text into labels based on meaningMarking a review as positive or negative, or an email as spam
SummarisationPulling out the key points of long textShortening a research paper or a news article
TranslationMoving text between languages while keeping the meaningTranslating a web page or an email
Content generationCreating new text or code from a promptSlogans, captions, stories, or a short Python script

If you want to go deeper on the prompt side, see our guide on how to write better prompts.

Create training content using Microsoft Copilot

This unit is a short video walkthrough. It introduces Microsoft Copilot (earlier called Bing Chat), an AI assistant that combines search, chat and content creation in one place. Copilot runs on an LLM, so it can answer questions in full sentences, and it also has a text-to-image feature that turns a description into a picture.

The module's example is teaching: using Copilot to draft lesson material, summarise a topic, and talk an idea through. The same approach works outside the classroom. If you run workshops, write guides or prepare team briefings, you can use an assistant to produce a first draft and then edit it yourself.

Tip: You do not need Microsoft Copilot to follow this module. The ideas apply to any LLM-based assistant, including ChatGPT and Claude. Our beginner AI tools list compares them.

Apply generative AI to enhance your training experience

Here the module lists where generative AI is already being used. It frames them for learning, but they apply to work in general:

  1. Content creation and design. Writers, designers and teachers can produce text, images and audio faster.
  2. Personalised marketing. Sales and marketing teams use it to tailor messages and improve engagement.
  3. Chatbots and virtual assistants. They answer customer questions quickly and can suggest products from the conversation. A personal assistant like Copilot can help with meeting notes, troubleshooting, reminders and email replies.
  4. Translation and language processing. For example, a browser can translate a web page into your language while you read, and the module notes Microsoft Edge supports over 70 languages.
  5. Accessibility. Tools such as Immersive Reader give reading support to learners with learning disabilities, and Speaker Coach in PowerPoint helps people practise presenting.
  6. Predictive analytics and forecasting. Schools can analyse past performance to flag learners at risk of dropping out, so teachers can step in early.

Apply critical thinking when using generative AI

This is the most important unit. AI models learn from large amounts of data from many sources, so their output can carry machine learning bias, meaning unfair or skewed results caused by problems in the training data. Critical thinking, the ability to analyse, evaluate and improve your own reasoning, is how you catch that.

The module breaks critical thinking into four skills:

  • Interpretation. Reading beyond the literal words to understand the meaning.
  • Analysis. Splitting something into parts and seeing how they relate.
  • Synthesis. Connecting two or more ideas, such as comparing viewpoints from different sources.
  • Evaluation. Judging how credible or important a piece of information is.

It then gives five habits for using AI responsibly:

  1. Check accuracy. Verify facts. You can ask the AI to name its sources, then open them and confirm they are current, reliable and reputable.
  2. Ask questions and get feedback. What is this for? Who is it for? How reliable is the information?
  3. Compare and contrast. Ask the same thing in different ways and see whether the answers agree.
  4. Read the content policies. Tool makers publish rules about what you can and cannot create. Follow them.
  5. Know the legal requirements. Rules about AI are changing, and in some settings you must disclose that you used it.

Why this matters for you: if you use AI for anything involving money, health or legal choices, verification is not optional. An AI can sound confident and still be wrong. Treat it as a fast first draft, not a final source.

Module assessment

The module ends with a short knowledge check on Microsoft Learn. You need to sign in and answer every question correctly to earn a pass mark on your Microsoft Learn profile, and we cannot take it for you. To prepare, see whether you can answer these five questions without scrolling up:

  1. What is the difference between generative AI and AI that only predicts?
  2. Name the two types of generative AI models the module describes.
  3. What four jobs can an LLM do with text?
  4. What are the four parts of critical thinking?
  5. Give two things you should do to check an AI answer before using it.

Summary

Generative AI creates new content from patterns it has learned, and LLMs are the technology behind most of the chat assistants you use. They can speed up writing, learning, translation and planning. They can also be biased or wrong, which is why the module's last message is to evaluate every output and know the tool's limits.

The module suggests further reading, including Microsoft's Fundamentals of generative AI training page and the Career Essentials in Generative AI path from Microsoft and LinkedIn.

Source: this article summarises Microsoft Learn's Introduction to generative AI - Exploring the basics in our own words. We are not affiliated with Microsoft. Microsoft and Copilot are trademarks of Microsoft.

Frequently asked questions

What is generative AI in simple words?

It is AI that learns patterns from large amounts of data and then creates new text, images or speech from your prompt, instead of only predicting or sorting existing data.

What is a large language model (LLM)?

An LLM is a neural network trained on a huge amount of text so it can hold a conversation, summarise, classify, translate, answer questions and write text or code. GPT is one example.

How long is the Microsoft Learn module?

Microsoft lists it as a beginner module of eight units that takes about 22 minutes, and it ends with a knowledge check you can take after signing in.

Do I need Microsoft Copilot to learn this?

No. The module uses Copilot for its examples, but the ideas apply to any LLM-based assistant such as ChatGPT or Claude.

Why do I need critical thinking when using AI?

AI output can be biased or wrong because of problems in its training data. Checking facts, asking questions, comparing answers and reading the content policies helps you use it responsibly.

Is this article official Microsoft content?

No. It summarises the public Microsoft Learn module in our own words and links to the original. We are not affiliated with Microsoft.

Penny the squirrel holding an acorn

Take your first step

Ready to start using AI, one small step at a time?

Join the free live AI Masterclass and see how to start using AI in your work, step by step. No experience needed.

Join the free AI Masterclass → Want hands-on practice first? See the AI Breakthrough