# Adult Language Practice Log

Use with [Practicing a Language with AI: A Drill Partner, Not a Certified Teacher](/articles/adult-language-practice-with-ai-limits).

## The rule

Attempt every sentence, translation, or answer yourself before the correction is shown. AI corrects
and explains; it does not produce your first attempt.

## Weekly drill log

| Day | Focus (vocab / grammar / roleplay) | My unaided attempt logged? (Y/N) | Errors AI flagged | Rule I need to review |
| --- | --- | --- | --- | --- |
|  |  |  |  |  |
|  |  |  |  |  |
|  |  |  |  |  |

## Human conversation log

AI drilling is preparation. This table is the part it cannot replace.

| Date | Conversation partner (exchange app / meetup / tutor) | Duration | What I got stuck on | What I want to drill next week |
| --- | --- | --- | --- | --- |
|  |  |  |  |  |
|  |  |  |  |  |

## Monthly review

- Did I book at least one real human conversation this month? If no, that is the gap to close before adding more AI drilling.
- Is the model correcting every error, or letting minor ones pass? If sessions feel unusually smooth, ask explicitly: "Correct every deviation from standard grammar, even small ones."
- Am I attempting sentences unaided before seeing a correction, or drifting into asking for the answer first?

## What this log is not

Not a proficiency certificate. Not a substitute for a placement test, an ACTFL or CEFR assessment, or
a qualified instructor's judgment of your speaking level. Use it to track practice consistency, not to
self-certify a level.