AI Foundations

Build a clear vocabulary for AI and see how it already appears in everyday products and workplaces.

• A short history of AI and why recent progress feels sudden
• AI, machine learning, deep learning, and generative AI compared
• Rule-based software versus systems that learn from data
• AI in search, email filtering, maps, and recommendations
• Common myths: AI as thinking, understanding, or always correct
• Reading AI headlines critically

Hands-on lab: Audit your own day: list ten tools you use, label which contain AI, and classify the type of AI in each.

Understand training data, patterns, and predictions through a simple model you build yourself.

• Supervised, unsupervised, and reinforcement learning at a high level
• Training data, labels, and why data quality drives results
• Classification versus prediction with everyday examples
• Overfitting explained with a simple analogy
• Measuring a model: accuracy and its limits
• How errors differ between false positives and false negatives

Hands-on lab: Train an image classifier in Google Teachable Machine with your own sample images, then deliberately feed it poor data and compare results.

Learn how language models turn text into predictions and why that shapes both their power and their errors.

• Tokens, context windows, and next-token prediction
• Pretraining and fine-tuning in plain language
• Why models produce fluent but incorrect statements
• Temperature and why the same prompt can give different answers
• Differences between ChatGPT, Claude, Gemini, and Copilot
• Knowledge cutoffs and tools that add live web search

Hands-on lab: Ask the same five questions to two assistants, compare answers side by side, and mark every claim you can verify and every claim you cannot.

Survey image, audio, and document generation and the practical questions each raises.

• How image generation models differ from text models
• AI summarization of documents, meetings, and spreadsheets
• Multimodal assistants that read images and files
• Authorship, copyright, and disclosure considerations
• Deepfakes and synthetic media: spotting and responding
• Choosing between a general assistant and a specialized tool

Hands-on lab: Upload a non-confidential document to an assistant, request a summary, then check the summary line by line against the source and log any omissions or inventions.

Examine how AI systems fail and build habits that reduce harm in everyday use.

• Sources of bias in data, design, and deployment
• Hallucination and how to detect it
• What happens to data you type into public tools
• Handling confidential, personal, and regulated information
• Human-in-the-loop review for consequential decisions
• Reading an organization's or tool's data policy

Hands-on lab: Run a bias probe: give an assistant paired prompts that differ only in a name or role, document differences, and write a short risk note with mitigations.

Turn the concepts into a practical decision framework for your own work or study.

• A simple fit test: task, data, risk, and reviewer
• Spotting tasks where AI saves time versus adds review burden
• Writing personal or team AI-use guidelines
• Staying current without chasing every new tool
• Explaining AI limits to colleagues, clients, or classmates
• Next steps for continued practice

Hands-on lab: Apply the fit test to ten tasks from your role, then present your top three candidates and the review step for each to the cohort.