AI courses and videos, in five steps

Shows how to ask ChatGPT a first question by typing or speaking, read the answer, and ask a follow-up to make it simpler, with a reminder to double-check important advice.

No experience needed · checked 2026-10-10

Explains in plain words how an AI chatbot is trained and how it builds a reply one word at a time, then gives four habits for getting better results.

No experience needed · checked 2026-10-10

Introduces generative AI and what sets it apart from earlier AI: it creates new text, images and other content instead of only analysing what already exists.

No experience needed · checked 2026-10-10

Explains in simple terms how chatbots like ChatGPT learn from large amounts of text to produce answers, and asks whether that amounts to real intelligence.

No experience needed · checked 2026-10-10

Goes through what today's generative AI tools can and cannot do well at this point in time, and where their limits are.

No experience needed · checked 2026-10-10

Explains how AI, machine learning, deep learning, foundation models and generative AI fit inside each other, and how the field developed from one to the next.

No experience needed · checked 2026-10-10

Traces about 70 years of AI, from early programs and expert systems through machine learning and deep learning to generative AI and AI agents.

No experience needed · checked 2026-10-10

The first module of Google's AI Essentials course: what AI and machine learning are, how generative AI works, and what these tools can and cannot do well.

No experience needed · checked 2026-10-10

Shows how saying what you need, who it is for and what kind of answer you want turns a vague question into a useful one, and how to adjust the reply without starting over.

No experience needed · checked 2026-10-10

Shows how to ask ChatGPT about a utility bill without sharing personal details, and how to recognise when an answer needs checking elsewhere.

No experience needed · checked 2026-10-10

Explains two common reasons an AI can sound sure and still be wrong (made-up facts and agreeing with you too easily) and four habits for checking its answers.

No experience needed · checked 2026-10-10

Covers discernment: judging whether what an AI assistant gives you, and how it behaves along the way, is accurate and useful for your task.

No experience needed · checked 2026-10-10

Explains what hallucination means in AI models, why it can happen, and ways to spot it in your own conversations.

No experience needed · checked 2026-10-10

Walks through creating a Claude project, adding a knowledge base, writing project instructions and sharing it with teammates so context carries across chats.

Some experience · checked 2026-10-10

Anthropic prompt engineers discuss what makes a good prompt engineer, how to refine prompts, honesty, personas and metaphors in prompts, and how prompting may change.

Some experience · checked 2026-10-10

A long, example-driven tour of how one AI researcher uses chat assistants: picking a model, when to use thinking modes, web search, file uploads, data analysis, voice and images.

Some experience · checked 2026-10-10

Andrew Ng shows how to use AI chat tools for sourced answers with web search and deep research, for feedback and writing, and for making simple websites and apps without code.

Some experience · checked 2026-10-10

Anthropic's applied AI team builds a prompt for a car insurance claim task step by step: task context, tone, background data, examples, XML structure, output format and prefill.

Technical · checked 2026-10-10

Introduces how students can use Gemini as a study helper to understand hard subjects, research and write papers, and test arguments, and how to use it responsibly.

No experience needed · checked 2026-10-10

An OpenAI marketer pulls product docs, research and meeting notes into a launch blog draft that follows a set template, then reviews and updates it as details change.

No experience needed · checked 2026-10-10

Shows how Khanmigo tutors students without handing them answers, helps teachers plan and differentiate lessons, and what it offers school and district leaders.

No experience needed · checked 2026-10-10

A short video for students and families on how AI works and how to think critically about it when using it for schoolwork.

No experience needed · checked 2026-10-10

Shows K-12 teachers how Gemini for Education adapts explanations with analogies and builds scaffolded materials, using a 5th-grade greenhouse effect lesson.

No experience needed · checked 2026-10-10

Shows how Copilot in Outlook drafts an email from a short prompt, answers questions about past emails, and checks grammar and tone before you send. Needs a Microsoft 365 plan with Copilot.

No experience needed · checked 2026-10-10

Lesson 1 of Google's AI Boost Bootcamp: where Gemini shows up across Google's work apps, and a four-part prompt formula (persona, task, format, context).

No experience needed · checked 2026-10-10

Shows Copilot making several changes to a budget spreadsheet in Excel and recalculating it, all with your approval, with Copilot finding the cells to adjust.

No experience needed · checked 2026-10-10

Shows students how to prompt Gemini as a tutor, research assistant or debate partner, and how to tailor its answers to their own learning style.

No experience needed · checked 2026-10-10

Walks through Copilot's Study and Learn agent, which guides students step by step with questions and practice activities instead of handing over answers, and how teachers can use it.

No experience needed · checked 2026-10-10

A Google course on using AI in daily work: how AI works, writing and refining prompts, brainstorming with Gemini, using it inside Google Workspace, and making it a habit.

No experience needed · checked 2026-10-10

A session for teachers on using AI responsibly in class: student data and privacy, copyright, classroom guardrails, misinformation, and bias in AI systems.

No experience needed · checked 2026-10-10

Four Anthropic staff with ties to education discuss benefits and risks of AI in learning, how students and teachers should approach it, and what is still uncertain.

No experience needed · checked 2026-10-10

A short walkthrough of ChatGPT Work: connecting your work tools and files, setting up scheduled tasks, and handing over marketing and data analysis jobs for ChatGPT to carry through.

Some experience · checked 2026-10-10

Shows Claude in Excel explaining formulas across tabs with cell-level citations and updating assumptions while keeping existing formulas intact.

Some experience · checked 2026-10-10

Uses Gemini, with NotebookLM, in the planning phase of a project to go from a blank page to a project plan.

Some experience · checked 2026-10-10

Demonstrates Copilot in Teams meetings: AI notes with decisions and action items, live interpretation between languages, and audio and video recaps for catching up.

Some experience · checked 2026-10-10

Microsoft's education product team demos the Teach Module in Microsoft 365 Copilot: creating lesson plans, quizzes and flashcards and adapting existing material.

Some experience · checked 2026-10-10

David Malan explains how Harvard's CS50 built AI tools, including a rubber duck tutor that guides students toward answers instead of giving them, and how the tools were rolled out.

Some experience · checked 2026-10-10

Explains retrieval-augmented generation (RAG): having a language model look up current, trusted sources before it answers, so answers are more up to date and less often wrong.

Some experience · checked 2026-10-10

Builds a simple AI agent with Gemini 3 Pro and Google's Agent Development Kit that can search the web, then runs it in a web interface and looks at how it reasons.

Technical · checked 2026-10-10

Members of Anthropic's Claude Code team show how they work with the coding agent: giving it goals instead of single tasks, running routines remotely, and letting it check its own work.

Technical · checked 2026-10-10

Anthropic engineers explain where the Model Context Protocol came from, remote MCP servers and registries, and how to use MCP together with the Claude API.

Technical · checked 2026-10-10

OpenAI staff show how to build long-running agents with the updated Agents SDK that inspect files, run commands and edit code across multi-step tasks.

Technical · checked 2026-10-10

Step-by-step setup of OpenAI's Codex coding agent: installing the command-line tool and VS Code extension, writing an Agents.md file, prompting patterns, and connecting tools via MCP.

Technical · checked 2026-10-10

Teaches spec-driven development: writing a project constitution and markdown specs so a coding agent builds what you asked for, including in an existing legacy codebase.

Technical · checked 2026-10-10

Ten lessons on AI agents: what they are, choosing a framework, design principles, tool use, agentic RAG, planning, multi-agent systems and deployment, with code samples.

Technical · checked 2026-10-10

Hands-on course on connecting a model to your own data with retrieval-augmented generation (embeddings, vector search), then linking tools and services through MCP servers.

Technical · checked 2026-10-10

A short visual introduction to how large language models predict the next word, how they are trained on huge amounts of text, and the basic idea behind transformers.

No experience needed · checked 2026-10-10

Builds up a neural network that recognizes handwritten digits, showing with animations what neurons, layers, weights and biases are and the math that links them.

Some experience · checked 2026-10-10

Visual walk-through of how data flows through a transformer: tokens become embedding vectors, attention and perceptron layers update them, and softmax gives next-word probabilities.

Some experience · checked 2026-10-10

Harvard CS50 lecture on generative AI and prompting, then decision trees, machine learning, deep learning, large language models and hallucinations.

Some experience · checked 2026-10-10

A long general-audience lecture on how models like ChatGPT are built: internet text, tokens, pretraining, turning a model into an assistant, reinforcement learning, and why they make things up.

Some experience · checked 2026-10-10

First lecture of MIT's 2026 intro to deep learning: perceptrons, activation functions, training with gradient descent and backpropagation, and how to limit overfitting.

Technical · checked 2026-10-10

First lecture of Stanford's CS336 (Percy Liang, Tatsunori Hashimoto): why build language models from scratch, the course plan, and how tokenizers such as BPE split text.

Technical · checked 2026-10-10

Andrej Karpathy writes a small GPT in PyTorch, building self-attention and the Transformer piece by piece and training it on the Tiny Shakespeare dataset.

Technical · checked 2026-10-10

Builds micrograd, a tiny automatic differentiation engine in Python, to show how backpropagation computes gradients, then uses it to train a small neural network.

Technical · checked 2026-10-10

Harvard's full CS50 AI course in one video: search, knowledge, uncertainty, optimization, machine learning, neural networks and language, with Python projects.

Technical · checked 2026-10-10