Understand the technology
From subject knowledge to a focused answer.
For May and November 2027. This lesson is for SL and HL students. Practise the familiar 2, 2, 4, 4, 6 and 12-mark sequence as a classroom exercise, with clear guidance on the changed 2027 assessment.
01 | KNOW THE ASSESSMENT, THEN PRACTISE THE SKILLS
The IB confirms that, from first assessment in 2027, the case study is assessed in Paper 1 at both SL and HL . The former standalone HL Paper 3 has been removed. Do not assume that its exact question sequence transfers unchanged.
This page uses a 30-mark classroom writing sequence: 2, 2, 4, 4, 6 and 12 marks , reflecting the familiar practice pattern discussed with your teacher. It is not a prediction, specimen paper or official markscheme.
All students can practise the shared-content questions. The 12-mark exercise develops extended reasoning; its inclusion here does not establish a 12-mark SL question. Follow the current specimen materials and the actual instructions for your level.
Official IB assessment update · First-assessment-2027 subject brief
02 | THREE THINGS EVERY ANSWER MUST MATCH
Before writing, identify the command word , the subject of the question and the available marks. Also check limits such as “two”, “one advantage” or “for Visionary Studios”.
The command tells you what kind of thinking to show. The subject tells you what to address. The marks signal the expected development, but they are not a universal sentence-count formula.
A technically correct paragraph about a different issue may not answer the question. Read the whole task before recalling a prepared explanation.
03 | COMMAND WORDS IN PRACTICAL LANGUAGE
| Command | What to do | A common mistake |
|---|---|---|
| Identify | Name or select the requested item precisely. | Writing a long unrelated explanation. |
| Define | Give a precise meaning, including distinguishing features. | Giving only an example. |
| Describe | Set out relevant features or stages. | Listing labels without saying what happens. |
| Explain | Show why or how through linked reasoning. | Repeating the claim without its cause. |
| Discuss | Develop a considered examination of relevant perspectives and arguments. | Listing unsupported pros and cons. |
Use your current course guide for authoritative command-term definitions. The same command can appear at different mark allocations depending on the task.
04 | IDENTIFY: BE PRECISE AND RESPECT THE NUMBER
If asked to identify two factors, give two distinct factors. Avoid several alternative guesses that make it unclear which answers you intend.
For example, “GPU memory capacity” and “available processing capability” identify different computing considerations. “A good computer” is much less precise.
Use the requested source where the question specifies one. You should not invent company facts to make a short answer sound more specific.
05 | DEFINE: MEANING BEFORE EXAMPLE
A definition should distinguish the term from nearby ideas. “Generative AI is AI that generates” simply repeats the wording.
A clearer definition says that generative AI learns patterns from data and uses them to produce new content, such as images. The example supports the meaning but does not replace it.
Definitions need not force the company name into every sentence. Apply the definition to the case when the question requires it.
06 | DESCRIBE: GIVE THE FEATURES OR STAGES
For a process question, organise the relevant stages in a sensible order. If describing image generation through iterative denoising, distinguish the starting noise, repeated updates and final output.
“Noise, model, image” supplies labels but says little about the process. Explain what changes at each stage while keeping the focus on what happens.
For a feature question, describe the relevant characteristics rather than forcing an artificial sequence. Always follow the subject of the task.
07 | EXPLAIN: BUILD THE CAUSAL LINK
An explanation connects a cause or mechanism to a consequence. “More retries take longer” becomes stronger when you explain that each rejected candidate requires another generation and review cycle.
Then connect that consequence to the company: total production time may increase, reducing the apparent speed advantage for client delivery.
Because, resulting in, leading to and therefore can help expose the link. They do not earn credit by themselves; the relationship must be technically sound.
08 | APPLY THE CASE THROUGH CONSEQUENCES
Adding “at Visionary Studios” to a generic statement is a weak form of application. Explain why the issue matters for the studio’s creative work.
Generic: “Consistency is important.” Applied: “If a recurring campaign character changes appearance between scenes, designers must correct the assets before client approval, increasing editing effort.”
Use details supplied by the booklet or question. Label any invented campaign as an illustration rather than an event that the studio actually experienced.
09 | DISCUSS: DEVELOP DIFFERENT PERSPECTIVES
A discussion explores relevant arguments, their consequences and the conditions affecting them. It can consider technical, operational and stakeholder perspectives where these answer the question.
For example, more generated concepts may help clients choose a direction, while extra review may increase designers’ workload. Explain both rather than count the number of advantages and disadvantages.
Finish with a supported synthesis where appropriate. Balance means considering significant competing arguments, not making every side equally convincing.
10 | USE MARKS TO PLAN DEPTH
| Practice task | Marks | Writing focus |
|---|---|---|
| Identify | 2 | Two precise distinct items when requested |
| Define | 2 | Accurate meaning with distinguishing detail |
| Explain | 4 | Developed causal reasoning |
| Describe | 4 | Relevant features or ordered stages |
| Discuss | 6 | Focused developed perspectives |
| Discuss | 12 | Sustained case-specific reasoning and synthesis |
Do not assume one sentence or paragraph equals one mark. Shorter items may use specific marking points; extended answers can be assessed through broader descriptors. Use the actual paper and markscheme when available.
11 | WORKED SHORT ANSWERS: 2 AND 2 MARKS
Identify two output characteristics relevant when evaluating images for Visionary Studios. [2]
Reveal an illustrative answer
Adherence to the requested visual style; preservation of required character identity across related images.
Why it fits: two distinct relevant characteristics are named directly.
Define dataset curation. [2]
Reveal an illustrative answer
Dataset curation is the deliberate selection, checking, organisation and documentation of data for a particular purpose.
Why it fits: it explains the activity and its purposeful nature, rather than merely naming a dataset.
12 | WORKED EXPLANATION: 4 MARKS
Explain why a short generation time does not necessarily mean faster delivery of an approved image for Visionary Studios. [4]
Reveal an illustrative answer
A generator may produce candidates quickly but fail the brief, so designers must make repeated requests and review the new results. Those cycles add production time. Selected candidates may also need editing to preserve character or product details before client approval, therefore the end-to-end delivery time can remain high despite a short individual generation step.
Why it fits: the answer develops retries and editing into consequences for delivery. This is illustrative reasoning, not a guaranteed four-mark allocation.
13 | WORKED DESCRIPTION: 4 MARKS
Describe how an image can be generated through iterative denoising in a diffusion-based system. [4]
Reveal an illustrative answer
The process starts from a noisy sample. At each generation step, a trained network provides information used to update the sample towards a less noisy state, guided by the conditioning input where applicable. Repeated scheduled updates develop image structure. The resulting representation is converted to a displayable image, including decoding if the process takes place in a latent representation.
Why it fits: it describes the progression without claiming that inference simply retrieves a stored original or retrains the model.
14 | WORKED DISCUSSION: 6 MARKS
Discuss the use of human review to improve the suitability of generated advertising images for Visionary Studios. [6]
Reveal an illustrative answer
Human reviewers can compare candidates with the client brief and identify inconsistent details or inappropriate portrayals. This helps the studio reject unsuitable assets before publication and allows contextual judgement that a single automated score may miss.
However, review requires staff time and reviewers can disagree or introduce their own biases. Checking only visual appearance also leaves other issues unresolved, such as source documentation or whether the client understands the AI contribution.
The studio should therefore define review responsibilities and criteria, use relevant expertise and escalate unresolved concerns. Human review is useful within a wider approval process, but it does not guarantee suitability on its own.
15 | PLAN THE EXTENDED RESPONSE BEFORE WRITING
For the 12-mark practice, plan a position, two or three important lines of argument, a significant limitation or alternative and a supported conclusion. This is a planning aid, not an official paragraph formula.
Each developed paragraph can connect a technical point to a company consequence, then qualify it where needed. Avoid a long introduction retelling the whole booklet.
If evidence is supplied, use it accurately. Do not import the fictional pilot figures from these lessons as if they were official case findings.
16 | WORKED EXTENDED DISCUSSION: 12 MARKS
Discuss whether Visionary Studios should use generative AI as its main method for producing final advertising images. [12]
Reveal an illustrative extended response
Generative AI could support the studio by producing alternative compositions and visual styles for client consideration. Exploring directions earlier may reduce effort spent developing unsuitable concepts. However, this potential benefit does not establish that generation should become the main method for final advertising assets.
Final production requires outputs that meet the actual brief. If a character or product changes between images, designers may need repeated generation and editing. Therefore the studio should compare total effort per accepted asset with an appropriate existing workflow, rather than rely on the speed of a single generation. A representative pilot should also inspect consistency across related scenes.
Creative flexibility is another opportunity. Generation can help communicate unusual concepts that would take time to stage or draw. Yet precise requested changes may alter other required details, so a human-led or combined workflow may provide better control for some commissions. Different production tasks may justify different methods.
Responsible use affects clients, creators and people represented in the images. The studio should document sources and relevant conditions, review portrayals and explain the AI contribution clearly. These actions support informed approval and trust, but they require staff time and do not automatically resolve every uncertainty. A visually impressive output is insufficient evidence that all concerns have been addressed.
Computing resources, provider access and staff skills also influence suitability. Training or adapting a system can require resources beyond ordinary generation, while hosted tools introduce dependency on service conditions. The studio should evaluate the workflow it actually intends to use and include setup, review and ongoing effort in its comparison.
I would recommend a staged adoption for suitable tasks rather than immediately making generation the main method for all final advertising images. Expansion should depend on evidence that representative assets meet agreed quality and consistency conditions with useful overall benefits, alongside completed approval checks. Retaining human-led alternatives allows the studio to use creative opportunities where demonstrated without assuming every task has the same solution.
Commentary: this answer weighs relevant opportunities and limitations, explains consequences and reaches a conditional judgement. It is an independent model response, not an official IB answer or promise of a particular score.
17 | SELF-CHECK THE REASONING
Read each answer against the exact question. Have you supplied the requested number of items, used precise terminology and distinguished a description from an explanation?
For longer answers, underline the company consequence in each paragraph. If none is present, consider whether the paragraph is too generic. Check that the conclusion follows from the arguments rather than introducing an unsupported claim.
Remove invented facts, repeated points and absolute promises such as “all risk is eliminated”. Check that examples illustrate the point and do not replace its explanation.
18 | PRACTISE UNDER REALISTIC CONDITIONS
Use the six worked prompts as a 30-mark classroom set: 2 + 2 + 4 + 4 + 6 + 12 = 30. Attempt them before opening the illustrative answers, then improve one weak response using a different colour.
Agree a practice time with your teacher. Plan proportionately around the marks and complexity, but use the actual duration and instructions for examination preparation rather than an assumed time for an old paper.
These shared-content exercises support SL and HL writing skills. Study HL-only technical material separately where required. Consult current official specimen papers and markschemes for the confirmed format and level of demand.
CHECK YOUR UNDERSTANDING | MATCH THE WRITING PURPOSE