Studying Ethically: Beyond the Rulebook
Starting university is a massive transition, and laying down a rigid list of "dos and don'ts" around ethics — especially with Generative AI — usually backfires. Students treat rules like compliance checkboxes to game or test, rather than a system of principles to live by.
To get incoming computer science and software engineering students to actually internalize ethical thinking, the narrative needs to shift from "What are the rules?" to "What kind of engineer or scholar do I want to be?"
The Dilemma of a Rules-Based Approach
Rules are inherently static, while software engineering and AI evolve dynamically.
| Fallacy | What it looks like |
|---|---|
| The Individual Fallacy | A rulebook tells you what you can get caught doing, not why you are doing it. If a rule says "do not use AI to generate code," students focus on evasion. If they understand that leaning on AI for fundamental logic deprives their brain of building the synaptic pathways for problem-solving, they realize they are only cheating their own future career. |
| The Societal Fallacy | In the real world, "it wasn't against the rules" or "the regulation didn't cover it" has led to dark patterns, algorithmic bias, autonomous safety failures, and privacy violations. Teaching students that compliance equals ethics creates engineers who build harmfully compliant systems. |
A Framework for Academic & Professional Integrity
Instead of rules, ground their studies in three classic ethical lenses made practical:
| Lens | The question it asks | Made practical |
|---|---|---|
| Virtue Ethics (focus on character) |
"Who am I becoming when I do this?" | If I prompt-engineer my way through my first-year programming modules, am I actually becoming a software engineer, or just a software consumer? |
| Utilitarianism (focus on outcomes) |
"What is the net impact of my action?" | When I push unverified, AI-generated code, am I creating systemic technical debt or security vulnerabilities that hurt end-users? |
| Deontology & Duty (focus on rights and fairness) |
"What if everyone did this?" | If every student outsourced their critical thinking to LLMs, what happens to the value of the degree, the trust of the industry, and the safety of the software powering society? |
Practical Application: The AI Continuum
Instead of "AI is banned" or "AI is allowed," present AI usage as a continuum of agency and learning outcome.
| Stage | What it looks like |
|---|---|
| [0] Zero AI | Writing code from scratch to build core mental models. |
| [1] AI as Tutor | Asking: "Explain this error message or concept to me like a mentor." |
| [2] AI as Co-Pilot | Debugging a syntax error after trying for 20 minutes. |
| [3] AI as Proxy — ❌ outsources the actual learning | "Write a class that does X, Y, and Z." |
The sweet spot for first-years sits between stages 1 and 2 — AI as a high-speed Socratic tutor, not a replacement for the struggle.
| When AI builds muscle | Using AI as a high-speed Socratic tutor. Asking it to explain a complex data structure, generate practice problems, or clarify an ambiguous error code. The student remains the active thinker; AI is the catalyst. |
| When AI causes atrophy | Using AI to bypass the "productive struggle." The pain of being stuck on a bug for an hour is where deep intuition is formed. Skipping that struggle means when systems break in production, the student won't have the mental models to fix them. |
Discussion Prompts for First-Year Workshops
To get students talking rather than listening to a lecture, split them into small groups with these scenarios:
-
The Short-Term Win vs. Long-Term Skill
"It's 11pm, your assignment is due at midnight, and your loop won't terminate. You prompt an LLM, it gives you working code, and you paste it in. You get an A. Did you win or lose?"
Facilitation angle: discuss how this affects their readiness for Year 2/3 and technical job interviews.
-
The "It's Not Explicitly Forbidden" Trap
"Your coursework brief doesn't explicitly mention AI usage for writing the reflection essay. Is using it ethical if there isn't a rule against it?"
Facilitation angle: transition from "rule checking" to virtue ethics — what does "doing original work" mean for an emerging scholar?
-
Scaling Up to Societal Impact
"You are asked by an employer to implement a feature that tricks users into subscribing to a service (a dark pattern). It's completely legal. Do you build it?"
Facilitation angle: show how the habits of personal academic integrity directly predict whether an engineer speaks up against unethical software in industry.
Key Takeaway Message for Students
"Rules exist to set minimum boundaries, but minimum boundaries do not make a great engineer. Computer science gives you the ability to build tools that shape how millions of people live, work, and think. Ethical engineering isn't about avoiding punishment from the university — it's about holding yourself to the standards required of someone who builds the digital infrastructure of tomorrow."