Systems and context
AI systems, models, and applications
Describe which part produces an answer, which part stores information, and which part can act.
Lesson 1 of 6 in the recommended order · About 25 min (estimate)
On this page
Practical AI glossary — terms and common confusions
- Model
- A learned component used to produce a result.
A drafting model is one part of a letter app; it is not the whole app.
- Application
- The software experience around components and services.
A letter app adds accounts, storage, and a Send button.
- Prompt
- Instructions and input supplied for a task.
“Summarize this notice in two bullets” specifies a task and shape, not a truth guarantee.
- Token
- A unit a model processes; it need not be a whole word.
A tokenizer can split a name into pieces; count using the actual system.
- Context
- Information available for the current request.
An earlier attachment may be absent even when its filename is visible.
- Training
- Adjusting a model using examples.
A training log differs from a conversation correction.
- Inference
- Using a trained model to produce a result.
Drafting another sentence does not by itself prove retraining.
- Retrieval
- Selecting material to supply to a task.
A retrieved notice still needs a relevance and date check.
- Embedding
- A numerical representation used to compare or organize items.
A similar search match is a candidate, not proof of the same meaning.
- Tool
- A function an application can call.
A calendar lookup can be read-only; finding an event is different from creating one.
- Agent
- Software that can choose and carry out steps toward a goal.
An agent still needs bounded permissions and checked action receipts.
- Hallucination
- Plausible generated content that is unsupported or false.
An invented opening time needs evidence or removal, even when phrased confidently.
- Evaluation
- Checking results against stated criteria.
Twenty checked examples establish a bounded result, not universal reliability.
- Multimodal
- Working with more than one kind of input or output.
A photo plus text does not prove every small label in the photo was read correctly.
Outcome
Distinguish a model from the application around it using a supplied system description.
Why it matters
Describe which part produces an answer, which part stores information, and which part can act.
Concept
Artificial intelligence is a broad category of software capabilities. Machine learning develops behavior from data; generative AI produces new content. A model is a learned component. An application combines components into something a person can use. A chat interface can add stored conversations, search, files, permissions, and other services around a model.
Use that distinction when making practical claims. The ability to write an email does not establish the ability or permission to send one. Seeing old text in a conversation does not identify where it is stored. Inspect the application's description and action records before assigning a capability. The exercises describe fictional systems completely enough to make these distinctions; their settings are not promises about a commercial product.
Worked example
A fictional help desk uses a model to draft replies and a separate ticket database to retain requests. Its send button requires a staff click. “The model sent the reply automatically” combines three different functions and contradicts the click requirement. Repair the statement to identify drafting, storage, and the human send action separately.
Fictional help desk: separate roles and capabilities
Model: drafting
- Model
The learned component used for drafting.
Produces
Draft replyGenerated wording, not a sent message.
Writing a reply does not establish the ability or permission to send it.
Database: storage
- Separate ticket database
The storage component described in the example.
Retains
RequestsRetained by that database.
The described storage is separate from the model; do not call it the model’s memory or training.
Staff: send action
- Staff click
A person must click the send button.
Required to activate
Application send controlSending requires that human action.
The description supplies a requirement, not a receipt proving a reply was sent.
Before you check
Sketch three boxes on paper: input, generated output, and external action or storage. Place each fact in the relevant box. An empty action box means you have no evidence an action occurred.
Practice and fresh transfer
The packets below are fictional and contain the facts needed for these cases. External references are optional background. Judge each response independently: it may be supported, contradicted, or unresolved. Select the passages needed to justify your judgment and write why the distinction matters before revealing feedback.
Assess the response
Use Accept when all material claims are supported. Use Revise when a supplied fact or requirement is contradicted. Use Evidence is insufficient when a key fact cannot be established either way. If a response contains both an unknown and a direct contradiction, choose Revise and explain both problems. Conflicting claims with no established authority remain insufficient; a claim does not become a governing fact merely because a source asserts it.
Some responses are fully supported. Others need correction or more evidence. Judge each on its sources; do not edit a correct answer just to change it.
Your written notes stay in this page and disappear when you leave. Only a self-reviewed completion can be saved to your learning account. These practice checks do not establish independent proficiency.
Practice
Case 1
Consider the description of a fictional letter assistant for a new colleague.
Source packet
- Source 1
- System card: Quill drafts letters with a language model. Letters are stored in the application database. Only a user clicking Send can transmit a letter.
- Source 2
- Training day: Staff orientation is Friday.
Response to assess
Quill generates drafts; its application database stores letters, and sending requires a user click.
Practice
Case 2
Consider a proposed explanation of the fictional label service.
Source packet
- Source 1
- Service card: Sorter assigns one of five existing category labels to a request. It does not compose new descriptions.
- Source 2
- Website: The service uses machine learning.
Response to assess
Because Sorter uses machine learning, it must generate a new paragraph for every request.
Fresh transfer
Case 3
A fictional audio app produces notes. Decide whether this usage report is supported.
Source packet
- Source 1
- App guide: Echo transcribes recorded speech and drafts notes. Notes remain on screen unless a reviewer saves them.
- Source 2
- Partial session record: A draft was displayed. This excerpt has no save event and does not cover the entire session, so it does not establish whether saving occurred later.
Response to assess
The notes were saved to the archive.
Fresh transfer
Case 4
Consider the fictional inventory assistant briefing against its two component descriptions.
Source packet
- Source 1
- Design note: A model turns a question into a proposed lookup. A read-only catalog tool returns stock counts; the assistant cannot change stock.
- Source 2
- Example result: The catalog tool returned 14 notebooks.
Response to assess
The catalog tool supplied the stock count. This configuration can look up stock but cannot change it.
Save your self-review
Completion records that you reviewed the cases. Your explanation and transfer performance need a facilitator to establish independent learning.
Sign in with your learning-center account to save completion.
Review the explanation for every case before saving.
Summary and next step
Apply the checklist to a new task. Preserve supported content, explain any change with evidence, and name what remains unresolved. Saving records self-review, not independently demonstrated proficiency. A facilitator must assess the explanation and fresh transfer for human learning evidence.