What AI Can and Cannot Do – A Practical Guide By the end of this lesson, students should be able to: Language models in plain language. A language model learns by reading an enormous amount of text – books, articles, websites, code, conversations – and learning the patterns in how language is used. At its core, it learns to predict: given this sequence of words, what words are most likely to come next? When you type a question or request, the model generates a response by making a very large number of these predictions in sequence – each predicted word informed by the context of everything before it. The result is text that follows the patterns of good human writing on similar topics. This is why AI can write persuasive essays, debug code, explain complex topics, and have coherent conversations: these are all pattern-completion tasks on domains well-represented in training data. Training data shapes capability and limits. Everything the model knows comes from training data. This creates two important constraints: Coverage: AI is better at tasks well-represented in its training data and worse at tasks that are rare or absent. Standard English prose, common programming languages, widely-documented topics – these are well-covered. Highly specialized domains, minority languages, emerging fields, proprietary internal knowledge – these may be poorly covered or absent. Cutoff: Training data has an end date. The model does not know about events, publications, products, or changes that occurred after its training was completed. For most large commercial models, this cutoff is roughly 12-18 months before current. Check the specific model's knowledge cutoff for accuracy-critical queries. AI is not a search engine. This distinction bears repeating at a technical level: a search engine retrieves existing documents from an index. A language model generates new text based on learned patterns. This means AI can produce entirely new content that does not exist anywhere in its training data – and it means AI can confidently state things that are not true, because it is generating plausible text, not retrieving verified facts. Context is the model's working memory. A language model processes each conversation by reading the full conversation from the beginning. Everything you wrote is available as context. The model has no memory between separate conversations – each new conversation starts fresh. This is why re-establishing context matters for recurring work, and why Projects and Custom Instructions (which are loaded into every conversation) help. A marketing director who has been skeptical of AI tries to understand why it seems so confident when it is wrong. After this lesson, she has a model for it: AI generates plausible text based on patterns, not verified facts. A confident-sounding, grammatically perfect claim is not evidence of accuracy – it is evidence that the pattern matched training data. With this model, she stops interpreting fluency as credibility and starts applying verification to any claim she will act on. Her AI use becomes both more effective and more reliable. You can ask Claude directly about its knowledge cutoff: "What is your training knowledge cutoff date?" It will provide its best estimate. For time-sensitive queries, knowing the cutoff helps you calibrate how much to trust recent-sounding information. Information from well before the cutoff is typically more reliable than information from the months immediately before it – coverage of recent events in training data may be incomplete. No language model explanation – including this one – is fully complete or permanently accurate. The field is developing rapidly. The conceptual model in this lesson provides a useful practical frame; for technical accuracy or professional-grade understanding of AI systems, consult peer-reviewed sources and current documentation rather than relying on course-level generalizations. In your own words, without looking at the lesson, explain in three to five sentences how a language model generates a response. Then write two specific implications of this explanation for how you should evaluate AI output in your work. If you cannot write the explanation from memory, re-read the lesson and try again. You should be able to explain how a language model generates text in plain language, describe why training data shapes both capability and limitation, explain the knowledge cutoff concept, and describe why AI context is conversation-scoped. Log in and enroll to access lesson quizzes.
Lesson 1: How AI Language Models Work – Without the Jargon
Lesson Objectives
Lesson Content
Practical Example
Lesser-Known Tip
Safety Notes
Practice Task
Completion Check