EngineerTeen: teach a robot to learn well

EngineerTeen: teach a robot to learn well

This guide was checked on 2026-10-19.

Editorial overview

The National Academy of Engineering launched EngineerGirl in 2001 and is expanding that work through EngineerTeen to reach a wider youth audience. The EngineerGirl Writing Contest began in 2003 and has used different topics and formats over its history. That background helps explain why a writing competition can ask such an engineering-minded question: what would a robot learn from the people around it, and who would take responsibility when it learned the wrong thing? The interesting work happens between imagination and explanation. A convincing robot is less important than a convincing account of how its behaviour changes.

CompeteMap sees this as a particularly useful crossover for students who enjoy telling stories but also question how technology affects ordinary life. A library, a rehearsal room or a familiar classroom can provide more persuasive material than a spectacular futuristic setting. The challenge is to notice an ambiguous situation, show why the robot misreads it, and propose guidance that actually addresses the cause. Writing and pictures need to do different jobs while telling the same story. We would prioritise a modest, carefully understood misunderstanding over an elaborate adventure in which the robot is simply told to be nice. The preparation burden is chiefly research, revision and ethical judgement, rather than expensive equipment or polished drawing technique.

Quick Facts

For current dates, eligibility and registration details, see the EngineerGirl / EngineerTeen Writing Contest competition page.

Review Evaluation

Rated Advanced. The challenge is connecting believable learning, ethical judgement and clear visual storytelling rather than building or programming a robot.

Begin with something the robot could misunderstand

A useful starting exercise is to observe a place the student already knows. Who speaks first? Who waits? Which rules are written down, and which are understood without explanation? A robot that merely copies visible behaviour could mistake silence for disinterest, speed for understanding or frequent speaking for having the best idea. Each offers a possible story, but the student should choose one they can explain precisely.

Take turn-taking in a reading group. A robot might repeatedly call on the quickest volunteer because that pattern appears to keep the activity moving. The story becomes interesting when another learner has something valuable to contribute but needs longer to respond. The engineering question is no longer whether the robot is friendly. It is what evidence it uses, what it fails to observe, and how someone can correct that failure without creating another unfair rule.

That is an invented planning example, not an official model answer. Its value is the chain of reasoning: observation, interpretation, consequence, feedback and changed behaviour. Students should be able to explain every link. If a proposed correction would not affect what the robot notices or how it decides, it is probably a moral slogan rather than a training strategy.

Research can sharpen the distinction between a system that follows a fixed instruction and one that learns patterns from examples. Students do not need to turn the narrative into a textbook. They do need enough understanding to avoid magical learning: a robot should not acquire a complex social judgement instantly just because the plot needs it to.

Let the storyboard carry the change

The storyboard is not a decoration attached after the writing is finished. Treat it as a test of whether the explanation has a visible sequence. The reader should see a situation, a mistaken interpretation, a response and an improvement. If every picture shows essentially the same smiling robot, the panels are not yet doing enough explanatory work.

A practical drafting method is to sketch tiny boxes before producing a polished image. Give each box one purpose. The student might show the robot noticing a repeated pattern, acting on that pattern, encountering a problem, receiving a corrective example and trying a different approach. The narrative can explain the uncertainty and the reasons for the correction; the pictures can make the sequence immediately understandable.

The official format requires a storyboard of 4–10 panels. Clear stick figures and simple symbols may communicate a useful idea better than detailed artwork with an unclear sequence. Draw attention to a decision: where the robot looks, whose contribution it misses, or what changes after feedback. Consistent characters and setting help the reader compare the situation before and after the correction.

Have someone unfamiliar with the draft describe what happened using only the panels. Then ask them to read the narrative. If the two accounts disagree, revise the connection rather than adding more visual polish. This is a communication problem, and finding the misunderstanding early is productive work.

Match the ambition to the grade band

The current rules distinguish four categories with different expectations. Use the equivalent US placement for an international or homeschooled student, rather than choosing a category because its word allowance looks more convenient.

The strongest older-student response will usually resist an effortless solution. A robot that remembers every conversation might become more useful while also creating a privacy problem. A rule that treats everyone identically can ignore meaningful differences in what people need. Students should make the tension understandable and explain why their proposed balance is defensible, rather than claiming to have eliminated every risk.

Check the current rules for reference-list requirements as well as the narrative allowance. A bibliography should make the research traceable, not display an impressive collection of sources that never influence the story. Keep notes about which explanation helped clarify a concept and which claim still needs checking. That habit also makes it easier to distinguish the student's reasoning from material encountered while researching.

Read the judging priorities as editorial advice

The published judging criteria examine AI learning, narrative voice, guidance strategies, ethical awareness and storyboard clarity. Engineering professionals participate in the volunteer judging panel, and the Pathways into Engineering Steering Committee judges finalists. Read these priorities together: technical understanding and personal voice need to support each other.

The official archive includes a past theme about engineering for communities and another about responsible engineering. Those examples help explain the programme's recurring interest in the relationship between engineering decisions and the people affected by them. They are context, not instructions to recycle an earlier entry. The current task needs its own setting, mechanism and argument.

During revision, ask three questions. Is the robot learning something specific? Does the proposed guidance respond to the actual misunderstanding? Does the student's own perspective survive the technical explanation? A narrative can be neatly organised and still feel impersonal if it never reveals what matters to its author. Equally, a heartfelt story can remain unconvincing if the supposed learning has no understandable cause.

Keep assistance separate from authorship

The current AI policy permits support with research, brainstorming, explanations, organisation and language checks, while requiring the submitted response to represent the student's own thinking and work. Asking a tool to produce the response and then making small changes is not a substitute for authorship. Students remain responsible for checking information and meeting citation requirements.

Adults can be useful readers without becoming co-authors. Ask where the explanation becomes confusing, what a panel is meant to show or why one correction would work better than another. Let the student make the decisions and write the answers. Keeping sketches and drafts is sensible practice because they show how the thinking developed.

Before submitting, check the official form with a parent or guardian. The narrative and storyboard need to be included in the submission itself; an external document link is not a replacement. Keep identifying information out of the work shown to judges, and leave enough time to check that both components are present and readable.

Key Takeaways

• Start with a real learning environment and one believable misunderstanding.
• Show how observation, feedback and changed behaviour connect; good intentions alone do not explain learning.
• Use the storyboard to reveal a sequence and the narrative to explain the decisions behind it.
• Check the appropriate category and current submission rules before polishing the work.
• The worthwhile outcome is a clearer understanding of how technology learns from people, including the mistakes and assumptions people pass on.

Sources checked

• https://www.engineerteen.org/4586/contest
• https://www.engineerteen.org/23559
• https://www.engineerteen.org/4585/About-EngineerTeen
• https://www.engineerteen.org/10143/Past%20Contests
• https://www.engineerteen.org/23757/2027-writing-contest-class-project-tips