Impacts and Ethics of Computing · Semester 2

Artificial Intelligence and Ethics

Discussing the benefits and risks of AI, including bias in machine learning models and accountability.

About This Topic

Artificial Intelligence and Ethics guides Secondary 4 students through the benefits and risks of AI systems. They explore how machine learning enhances fields like healthcare diagnostics and traffic management, while addressing dangers such as bias in models that favor certain demographics. Discussions center on accountability for errors in autonomous systems, methods to build fair algorithms, and AI's influence on future jobs and creative processes.

This topic aligns with MOE Computing and Society standards for S4, as well as Artificial Intelligence objectives. Students apply ethical frameworks to evaluate real Singaporean contexts, such as AI in public housing allocation or national service predictions. These connections build skills in critical analysis and responsible tech citizenship.

Active learning proves essential for this abstract topic. Role-playing ethical dilemmas or debating case studies in small groups helps students confront biases firsthand and weigh accountability trade-offs. Such approaches make ethics personal and actionable, deepening retention and preparing students for informed societal contributions.

Learning Objectives

• Analyze case studies of AI implementation in Singapore to identify potential ethical risks such as algorithmic bias or lack of accountability.
• Evaluate proposed solutions for mitigating bias in machine learning models, comparing their effectiveness and feasibility.
• Critique the societal impact of AI on employment and creativity, synthesizing arguments for both positive and negative transformations.
• Design a set of ethical guidelines for the development of a hypothetical AI system, considering principles of fairness, transparency, and accountability.

Before You Start

Introduction to Artificial Intelligence

Why: Students need a basic understanding of what AI is and its common applications before exploring its ethical implications.

Data Representation and Analysis

Why: Understanding how data is collected and analyzed is fundamental to grasping concepts like algorithmic bias and fairness in machine learning.

Key Vocabulary

Algorithmic Bias	Systematic and repeatable errors in a computer system that create unfair outcomes, such as privileging one arbitrary group of users over others.
Accountability	The obligation of an individual or organization to accept responsibility for their actions and decisions, especially when autonomous systems cause harm.
Transparency	The principle that the workings of an AI system, particularly its decision-making processes, should be understandable and explainable to users and stakeholders.
Machine Learning	A type of artificial intelligence that allows systems to automatically learn and improve from experience without being explicitly programmed, often by identifying patterns in data.

Active Learning Ideas

Debate Rounds: AI Accountability

Assign small groups to roles: developers, users, regulators. Provide cases like self-driving car accidents. Each group prepares 3 arguments in 10 minutes, debates in rounds of 4 minutes per side, then votes on resolutions. End with individual reflections on key takeaways.

⏱ 45 min·Small Groups

Generate mission

Bias Detection Lab: Dataset Scrutiny

Distribute sample datasets on loan approvals or facial recognition. Pairs identify biases by charting demographics and error rates. Groups propose debiasing steps, such as data augmentation, and share via class gallery walk.

⏱ 35 min·Pairs

Generate mission

Ethical Dilemma Cards: Role Play

Deal scenario cards on AI in hiring or creative arts. Small groups role-play stakeholders discussing solutions for 15 minutes. Perform skits for class, followed by whole-class criteria for ethical AI.

⏱ 50 min·Small Groups

Generate mission

Future Work Vision Boards

Individuals brainstorm AI impacts on 5 jobs, then pairs create vision boards with pros, cons, and adaptations. Share in whole-class carousel for collective insights on reskilling needs.

⏱ 40 min·Individual

Generate mission

Real-World Connections

In Singapore, AI is used in the public housing application process to help allocate flats. Students can explore if the algorithms used are fair and do not inadvertently discriminate against certain demographics.

The Singapore Police Force is exploring AI for predictive policing. This raises questions about accountability if an AI wrongly identifies a suspect or if bias leads to over-policing in certain neighborhoods.

Watch Out for These Misconceptions

Common Misconception

AI is neutral and unbiased by design.

What to Teach Instead

AI inherits biases from training data that mirrors societal inequalities. Small-group dataset analyses reveal patterns, like underrepresentation of minorities, helping students grasp data's role and test fairness metrics collaboratively.

Common Misconception

No one is accountable for AI errors since machines decide independently.

What to Teach Instead

Responsibility traces to designers, deployers, and overseers. Role-playing chains of decisions clarifies this, as students negotiate outcomes and refine their views through peer feedback.

Common Misconception

AI will replace all human jobs completely.

What to Teach Instead

AI often augments roles, creating new opportunities alongside changes. Debates on real cases, such as AI in Singapore's finance sector, show hybrid models, with groups mapping transitions to build nuanced predictions.

Assessment Ideas

Discussion Prompt

Present students with the scenario: 'An autonomous vehicle causes an accident resulting in injury. Who is responsible: the programmer, the owner, the manufacturer, or the AI itself?' Facilitate a class debate where students must justify their assigned role's accountability using ethical principles.

Quick Check

Provide students with a short description of a machine learning model used for loan applications. Ask them to identify one potential source of bias in the data used and suggest one method to mitigate it. Collect responses to gauge understanding of bias and mitigation strategies.

Exit Ticket

Ask students to write down one way AI might change a job they are interested in, and one ethical concern related to that change. This helps them connect AI's future impact to personal aspirations and ethical considerations.

Frequently Asked Questions

How can teachers address AI bias in Secondary 4 Computing lessons?

Start with relatable Singapore examples, like biased facial recognition in MRT security. Use hands-on dataset audits where students quantify imbalances and brainstorm fixes, such as diverse training sets. Follow with discussions linking bias to ethical standards, reinforcing MOE goals for fair computing. This builds analytical skills through concrete evidence.

Who is responsible when AI makes a harmful decision?

Accountability spreads across developers for flawed algorithms, organizations for deployment, and regulators for oversight. Students explore this via cases like algorithmic lending errors. Structured debates help them map responsibilities, emphasizing transparency requirements under Singapore's PDPA and emerging AI governance.

How can active learning help students grasp AI ethics?

Active methods like role plays and bias hunts engage students directly with dilemmas, turning abstract concepts into lived experiences. Groups debating accountability or auditing datasets uncover nuances collaboratively, improving critical thinking over lectures. Reflections solidify connections to real impacts, aligning with MOE's emphasis on student-centered inquiry for deeper ethical understanding.

What ethical issues arise from AI redefining work and creativity?

AI automates routine tasks but raises concerns over job displacement and authorship in generated art. In Singapore's context, consider AI in creative industries like media. Class visions boards prompt students to weigh augmentation benefits against reskilling needs, fostering proactive ethical stances on workforce evolution.