A quick draft is only one part of a finished task. If you must repair details or spend longer checking, the first output’s speed may not help. The calculation below lets you practice that distinction without buying a service or experimenting on student work.
Define the finished task first
For an invented notice, completion means the facts and formatting have been checked, not merely that text appeared. Compare like with like: the same requirements, checking standard and unit of time. Keep one-time learning effort visible rather than quietly omitting it.
NIST’s guidance describes risks from false output and overreliance. It does not establish how many hours a teacher will save. This page withdraws the old unsupported workload percentages and named-tool performance assurances; it offers no replacement forecast.
Include work that the first draft hides
A useful record can separate active work from waiting. State which total you mean. Do not add overlapping intervals twice. If a result fails its factual checks, do not count it as a completed equivalent task just because it was fast. For this practice, all intervals are sequential, so the arithmetic is straightforward.
Practice: compare two fictional workflows
Supplied facts: A fictional adult-club notice takes 20 minutes to write and check manually. An alternative takes 4 minutes to prepare instructions, 2 minutes waiting, 8 minutes checking, and 7 minutes correcting and checking again. A separate introductory tutorial takes 10 minutes. All figures are invented, sequential intervals; both final notices meet the same supplied requirements.
Practice prompt: “Calculate active effort, elapsed task time and first-use total including the tutorial. Compare each relevant total with the 20-minute manual task. Do not predict a weekly saving.” No tool is needed beyond paper or a calculator.
Deliberately flawed sample: “The alternative takes 4 minutes, saving 16 minutes per notice and hours every week. Checking does not count because the draft is already finished.” This is an authored calculation error.
One possible correction: “Active effort is 4 + 8 + 7 = 19 minutes. Elapsed task time is 19 + 2 = 21 minutes. Including the separate tutorial, first-use elapsed time is 21 + 10 = 31 minutes. Compared with the 20-minute manual task, active effort is one minute lower, task elapsed time is one minute higher, and first-use elapsed time is 11 minutes higher. No weekly saving is established.”
Decide from a complete comparison
Checks: Include checking and corrections. Keep waiting separate when reporting active effort, but include it in sequential elapsed time. State whether the tutorial is included. Reject a 16-minute saving or any weekly prediction: the supplied record does not support either. Manual drafting is a reasonable choice under these facts.
Limits: These numbers measure no real provider, workload or learner. Actual tasks may overlap, differ in difficulty or fail quality checks. This exercise provides no justification for automated grading, student uploads or reduced professional review. For privacy boundaries, read starting safely.
Continue with general AI practice
Open free Module 1. An account is required, with no credit card. The course overview describes the foundation material. External AI services have separate accounts, terms and charges; none is required for this exercise.
The learning-preferences quiz leads to a general rule-based plan using your goal, main obstacle and available time. Saving the plan asks for an email; it does not create a lesson account or email the plan. It does not assess teaching competence, approve classroom software or certify professional development.
For another exercise, read drafts and decisions or evidence and uncertainty.
Browse the teacher topic hub for related reading. Materially revised October 12, 2026 with assistant source checking. The exercise and sample answers are authored fiction, not measured AI outputs. This is not human teacher review, legal advice, a district approval or evidence of learning gains.
Frequently asked questions
- Will AI halve my workload?
- This page provides no evidence for that promise. Compare the whole task, including setup, checking and correction, before making any claim.
- Should I try automated grading on real student work?
- Not for this exercise. It uses a fictional adult-club notice and invented time records, with no grades or student information.
- Do faster drafts prove better teaching?
- No. Speed alone does not establish accuracy, accessibility, fairness or learning benefits.
- What if checking takes longer than manual work?
- That is a useful comparison result. You can choose the manual approach rather than add a tool to this task.
- Are the timing figures measured results?
- No. They are invented arithmetic inputs. They describe no teacher, product trial or expected completion time.
