Independent Study: Ethics, Human-AI Interaction, and Automated Reasoning in Scientific Domains

Course Overview

This independent study explores the ethical dimensions of computer-human interaction and algorithmic decision-making, while integrating a technical lens on automated reasoning and formal methods in AI. The student will investigate how ethical concerns, such as bias, opacity, and manipulation, intersect with algorithmic systems, and explore tools and methods to enhance AI's trustworthiness. Inspired by Fairness, Accountability and Transparency in AI and Automated Systems.

Learning Goals

• Analyze critical works on ethics in AI, surveillance, fairness, and human-computer interaction.
• Explore technical applications of automated reasoning.
• Evaluate tools to improve AI systems' reliability, interpretability, and safety in scientific contexts.
• Develop a reflective and practical understanding of ethical system design.

Outline

Foundational Knowledge

1. Define key ethical theories (e.g., consequentialism, deontology, virtue ethics) and explain how they apply to the design and deployment of intelligent systems.
2. Identify core governance frameworks, policies, and standards shaping AI development at national, regional, and global levels.
3. Describe major ethical challenges in intelligent systems, such as bias, transparency, accountability, privacy, safety, and sustainability.
4. Compare approaches to AI regulation and governance across jurisdictions (e.g., EU AI Act, U.S. AI executive orders, UNESCO AI Ethics guidelines).

Analytical & Critical Thinking Skills

1. Critically evaluate case studies of intelligent systems for potential ethical risks, unintended consequences, and governance gaps.
2. Assess the trade-offs between innovation, social good, and risk mitigation in AI policy and system design.
3. Analyze stakeholder perspectives (governments, corporations, communities, and marginalized groups) to understand competing values in AI governance.
4. Apply risk assessment tools and frameworks to identify and mitigate harms in AI lifecycle stages (data, models, deployment).

Practical & Applied Skills

1. Develop guidelines or audit checklists for ethical evaluation of an AI system in real-world contexts.
2. Draft governance recommendations that balance ethical principles with technical feasibility and organizational constraints.
3. Propose participatory approaches for involving diverse stakeholders in the oversight of intelligent systems.
4. Use fairness, accountability, and transparency (FAccT) metrics to evaluate AI models, where applicable.

Integration & Synthesis

1. Design a policy brief or position paper articulating an ethical stance on a contemporary AI issue (e.g., facial recognition, autonomous weapons, healthcare AI).
2. Formulate strategies for embedding ethical reasoning into the development cycle of intelligent systems.
3. Synthesize ethical theory, governance mechanisms, and technical methods to propose holistic solutions to real-world challenges.
4. Reflect on the societal and personal responsibilities of computer scientists and engineers working in AI.

Weekly Breakdown

Major Projects

Midterm Technical-Ethical Analysis

Analyze an existing biomedical or scientific reasoning system. Explore its:

• Representational logic (e.g., OWL, RDF, first-order logic)
• Ethical implications (bias, opacity, misuse)
• Formal guarantees (if any)

Deliverable: Ethical report

Final Project: Thesis Paper

Deliverables:

• Report or write-up (expecting this to be substantial in length, more details to be proposed as we zero in on a topic)
• 10-minute presentation