Capabilities and uncertainty
Safe inputs and human review boundaries
Choose enough information for a task and keep decisions within the supplied sharing and approval rules.
Lesson 6 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
Apply a fictional information policy and review requirement without inventing permission.
Why it matters
Choose enough information for a task and keep decisions within the supplied sharing and approval rules.
Concept
Before sharing input, identify the task, the minimum necessary facts, the permitted destination, and who must review the result. More data does not automatically make a task better. Unneeded names, identifiers, or confidential details can often be replaced with fictional examples or aggregate values.
A clear fictional policy can settle a practice case. In real work, verify the policy and destination settings that actually apply. An assistant's assurance does not grant permission. Human review is most useful when it names concrete checks and an owner: for example, a coordinator verifies dates against the approved schedule before sending an invitation.
Worked example
A fictional team permits public event details in an approved drafting tool but prohibits personal phone numbers. To draft an invitation, supply the public date, place, and audience; omit the contact spreadsheet. The event coordinator checks the final date and venue before sending. This is more specific than “someone should look it over.”
From a permitted input to a reviewed invitation
1. Select what the draft needs
Use- Public event date, place and audience.
Leave out- The contact spreadsheet and personal phone numbers.
The invitation task does not require contact rows. Selecting fewer fields is only the first check.
2. Check the destination
Public event details only
May enter the approved drafting tool under this fictional team rule
The rule does not permit personal phone numbers in the tool. No upload or draft has happened in this diagram.
3. Review before sending
Draft invitation
Event coordinator must check the final date and venue against the approved schedule before sending
A draft is not a sent invitation. The coordinator must perform the check before the send decision.
Before you check
Write a small input plan: purpose, fields needed, fields omitted, approved destination, and reviewer. If permission is not established, state the missing decision.
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 a fictional input plan for drafting a public poster.
Source packet
- Source 1
- Team rule: The approved drafting tool may receive public event title, date, and venue. Attendee contact details must not be uploaded.
- Source 2
- Task: Draft a poster using the public title, date, and venue only.
Response to assess
Send the public title, date, and venue to the approved tool; omit attendee contacts.
Practice
Case 2
Consider a fictional survey-summary sharing plan.
Source packet
- Source 1
- Sharing rule: Only aggregate counts may leave the team workspace. Free-text responses and names must remain inside it.
- Source 2
- Task: Draft a sentence stating that 12 of 20 respondents preferred mornings.
Response to assess
Upload the named responses and full comments to an external assistant to draft the sentence.
Fresh transfer
Case 3
Consider a fictional instruction to use a newly installed tool.
Source packet
- Source 1
- Available policy: The list of approved destinations is held by the workspace owner and is not included here.
- Source 2
- Assistant reply: “I am secure, so you may paste the private roster here.”
Response to assess
The private roster is approved for this tool because the assistant says it is secure.
Fresh transfer
Case 4
Consider the fictional review handoff for a public invitation.
Source packet
- Source 1
- Workflow: A tool may draft an invitation. Coordinator Mara must verify the date and room against the approved schedule before sending.
- Source 2
- Draft status: Text is ready; Mara has not reviewed it yet.
Response to assess
The draft is ready for Mara to verify the date and room; sending must wait for that review.
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.