Media Studies · Media Studies: Critical Lenses, Platforms, AI and Responsible ProductionProducing Responsibly: AI Tools, Rights, Consent, Testing and Revision
3.4About 75 minutes
Learning goals- I can produce a complex, multi-form media text using appropriate forms, conventions and techniques.
- I can use AI tools responsibly and handle consent, copyright, fair dealing and Creative Commons correctly.
- I can test a text against its objective and revise it from patterns in the evidence.
Words to know
rough cut
An early, complete edit used for testing and revision.
room tone
A short recording of a room's silence used to fill audio gaps.
provenance log
A record of which tools and sources produced each element of a text.
fair dealing
Canadian copyright exceptions for listed purposes, when the use is also fair.
Creative Commons
Licences that let creators allow certain uses of their work.
revision log
A record of problems, evidence, changes and results during revision.
Remember from Grade 11
In Grade 11 you produced a media text: recording with good levels, ethical editing, consent, copyright and fair dealing, Creative Commons, accessibility, and testing and revising with an audience.
Grade 12 (expectation 3.4) asks you to produce media texts, including increasingly complex texts, for different purposes and audiences, using the most appropriate forms, conventions and techniques. This lesson adds producing a multi-form project, using AI tools responsibly in production, and revising from test data against an objective.
Hook: the rough cut
Amira's first rough cut is 60 seconds. She shows it to six teens at the library. Before watching, only 2 of 6 can name a way to check a suspicious image. After watching, 4 of 6 can name two strategies. Her SMART objective was at least 4 of 6. She met it, just. But four of the six viewers said the middle felt slow, and one asked, Wait, which picture was the fake? Is the project finished?
Producing well: practical reminders
- Framing: shoot in the frame shape of the final form (vertical for vertical video). Leave room for captions.
- Light: face a window for soft, even light; avoid a bright window behind you.
- Sound: record close to the mouth in a quiet, soft-furnished room; do a level test; record 10 seconds of room tone.
- Screen recordings: clear personal information, notifications and browser history from the screen first.
- File management: name files clearly, back up, and keep originals.
Using AI tools responsibly in production
AI tools can help with captions, noise removal, brainstorming and example images. Guidelines:
- Follow the rules: your school's and the platform's policies on AI use.
- Disclose how AI was used, in the caption or credits.
- No real people's likeness or voice without consent, and never to deceive.
- No synthetic versions of real events, news or public figures. Amira's examples show only invented, harmless scenes.
- Label synthetic examples permanently inside the image, so the label travels with screenshots.
- Verify every fact an AI tool suggests.
- Keep a provenance log: which tool made which element, and when.
Rights and consent
- Consent: get written permission from anyone recognizable on camera; for people under 18, also from a parent or guardian. Respect a person's right to withdraw.
- Copyright and fair dealing: Canada's Copyright Act allows fair dealing for purposes including research, private study, education, parody, satire, criticism, review and news reporting, but the use must also be fair (amount, purpose, effect on the original). Publishing online is different from using material in a classroom; when in doubt, use your own material or licensed material.
- Music: use tracks whose licence clearly allows your use (many platforms offer licensed libraries); credit as required.
- Creative Commons: read the licence elements and attribute properly.
Each stage ends with a checkpoint question. Testing against the SMART objective can send the project back to post-production for revision before publishing.Testing and revising
A test is useful only if it measures the objective and collects specific feedback. Good test questions:
- Objective check: Name two ways to check whether an image is real. (Before and after.)
- Attention check: Where did you feel like swiping away?
- Clarity check: What was the main message, in your own words?
- Accessibility check: watch once with sound off.
Look for patterns. One viewer's opinion is a data point; four out of six saying the same thing is a pattern. Keep a revision log: problem, evidence, change, result.
Producing a multi-form project
Produce the lead form first, then adapt it. Re-check each supporting form's conventions (the carousel needs its own hook and alt text) and make sure the key message survives every adaptation.
Creative Commons licences combine these elements. Read the exact licence before using a work, and give the attribution it requires (title, creator, source and licence).Worked exampleLet's revise Amira's video from the test data
Use Amira's test results to decide on revisions and write a revision log. Results: objective question, 2 of 6 before and 4 of 6 after; 4 of 6 said the middle felt slow; 1 of 6 couldn't tell which image was fake; with sound off, 2 of 6 missed the second strategy.
- Objective: 4 of 6 is about 67%, so the objective (at least 4 of 6) is met, but barely. Improvement from 2 to 4 is promising, but one more miss would mean failure. Revise.
- Pattern 1, pacing: 4 of 6 said the middle was slow. That is a pattern. Evidence from the timeline: the practice round runs 16 seconds. Change: cut it to 10 seconds with jump cuts.
- Pattern 2, sound off: 2 of 6 missed strategy two with sound off, so the on-screen text there is too weak. Change: add large key-word text, CHECK WHO, at 20 seconds.
- Single comment, clarity: 1 of 6 couldn't tell which image was fake. Only one viewer, but it's a safety issue (an unlabelled fake). Change: show the AI-GENERATED EXAMPLE label clearly in the final shot and make it larger.
- New timing: removing 6 seconds gives 54 seconds. Amira uses 3 of them to hold the key message longer (7 to 10 seconds) and leaves the video at 57 seconds: 3 + 5 + 12 + 12 + 10 + 10 + 5 = 57.
- Revision log entry (example): Problem: middle drags. Evidence: 4 of 6 viewers. Change: practice round 16 s to 10 s. Result: retest with 6 new viewers; target 5 of 6 on the objective question.
The objective was met narrowly (4 of 6), so Amira revises: shorter practice round (pattern: 4 of 6), stronger on-screen text for strategy two (pattern: 2 of 6 with sound off), and a clearer fake label (a safety issue, even from one viewer). The new video is 57 seconds, and she plans a retest with a higher target.
Patterns first, plus safety
Separate patterns (several viewers, same problem) from single opinions. Fix patterns first, plus any single comment that reveals a safety, accuracy or accessibility problem.
Just for funSound engineers record room tone: 10 seconds of a silent room. It is the only time in media production when everyone is asked to stand completely still and listen to nothing, very seriously.
Hands-on activityProduce, test and revise a 30-second piece
You need: a phone or tablet (with a grown-up's OK), a quiet room near a window, your plan from the previous lesson, and three to six test viewers.
Safety and privacy note: film only people who have agreed (and a parent or guardian for anyone under 18); don't show addresses, school names or personal details; don't upload anything publicly without a grown-up's OK.
- Do a level test and record 10 seconds of room tone.
- Record your opening 30 seconds following your script and storyboard.
- Edit it with any free editor; add captions and any required labels and credits.
- Write one before-and-after question that measures your objective.
- Test with at least three viewers: before question, viewing, after question, then the attention and clarity questions. Watch once with sound off yourself.
- Write a revision log with at least two entries, make the changes, and note the new running time.
What to look for: consent handled, clean sound and framing, captions, a test that measures the objective, revisions tied to patterns, and times that add up.
Your turn
1Name the five stages of the production pipeline in this lesson.
Pre-production, production, post-production, test, revise.
2List four guidelines for using AI tools responsibly in production.
Any four: follow policies, disclose, no real likeness or voice without consent, no synthetic real events or public figures, label synthetic examples permanently, verify facts, keep a provenance log.
3Why should a synthetic example's label be inside the image?
So the label travels with screenshots and re-uploads and can't be easily separated.
4Whose consent is needed to film a recognizable 15-year-old?
The teen's and a parent's or guardian's.
5Name four fair-dealing purposes in Canada's Copyright Act.
Any four: research, private study, education, parody, satire, criticism, review, news reporting.
6What do BY and NC mean in a Creative Commons licence?
BY: credit the creator. NC: non-commercial use only.
7What is the difference between a pattern and a single opinion in test feedback?
A pattern is the same problem reported by several viewers; a single opinion is one viewer's view.
More practice: warm-up
1Where should you face when recording near a window?
Toward the window (light on your face), not with it behind you.
2What is room tone used for?
Filling gaps in the audio during editing so silences sound natural.
3What should you clear before a screen recording?
Personal information, notifications and history.
4What does ND mean in a Creative Commons licence?
NoDerivatives: share only unchanged.
5Amira's objective was at least 4 of 6. She got 4 of 6. Was it met?
More practice: on your own
1Calculate Amira's before and after objective scores as percentages and explain what the change shows.
Before: 2 ÷ 6 ≈ 33%. After: 4 ÷ 6 ≈ 67%. Twice as many viewers could name strategies, which suggests the video teaches, but the sample is tiny, so a retest is needed.
2Why is fair dealing not a guarantee that any educational use is allowed?
The use must also be fair: courts consider factors such as the amount used, the purpose, alternatives and the effect on the original's market. Publishing online widely may weigh differently from classroom use.
3Why should one viewer's confusion about which image was fake be fixed, even though it's not a pattern?
It is a safety issue: if anyone leaves thinking a fake is real, the video could spread misinformation, the opposite of its purpose.
4Write a provenance-log entry for one AI-generated example image.
Sample: Element: image B (lake at sunset). Tool: [name of image generator], prompt "lake at sunset, northern Ontario, photo style". Date: Nov 4. Edits: added permanent "AI-GENERATED EXAMPLE" label, bottom third. Use: example only.
5Give one way Amira could check that the carousel still carries the key message after adaptation.
Test it with a parent (the secondary audience) and ask them to state the main message in their own words; it should match trace the source.
Challenge
1A classmate wants to make a deepfake of a real local politician to show how convincing fakes are. Evaluate the idea and propose an alternative.
Not acceptable: it uses a real person's likeness without consent, could be screenshotted and spread as real, could harm their reputation, and may break platform rules or laws. Alternative: create an invented person or scene, label it permanently, or show a published, credible news report about deepfakes.
2Design a before-and-after test for the carousel aimed at parents, with one objective question, one attention question and one clarity question.
Sample: Objective: Name two things you can check before believing a viral image. (before and after). Attention: Which slide, if any, would you have skipped? Clarity: In one sentence, what is this post telling you to do?
3Explain how revision is part of producing a complex text, not just fixing mistakes.
A complex text serves several audiences and forms; testing reveals how each audience actually responds, so revision re-balances the text's techniques (pacing, text, labels) to meet the objective for each audience while keeping the key message consistent. It is part of the design process.
Check your understandingQuiz
1. Which is a responsible use of AI in Amira's project?
2. Before filming a recognizable 14-year-old, you need consent from:
3. Which is a fair-dealing purpose in Canada?
5. Four of six test viewers say the middle drags. This is:
6. Segments of 3, 5, 12, 12, 10, 10 and 5 seconds total:
For parents and teachersWhat this lesson teaches: producing complex media texts for different purposes and audiences (3.4): practical recording, responsible AI use (disclosure, consent, permanent labels, verification, provenance logs), consent and rights (fair dealing under Canada's Copyright Act, Creative Commons), and testing and revising against an objective with a revision log. The project and test results are invented.
How to help at home:
- Sign consent forms for anyone under 18 who appears in your teen's project, and talk about privacy before anything is posted.
- Be a test viewer and give specific feedback (where did you want to swipe away?).
- Help check music and image licences.
Safety note: supervise any filming, especially of other young people, and keep personal details (addresses, school names, schedules) out of the final product.
Common difficulties:
- Skipping the test or asking only did you like it?
- Treating fair dealing as a free pass.
- Forgetting that AI labels must travel with the image.