Ethics of Artificial Intelligence

Notes

• Course taught by Christian Tarsney (christian.tarsney@utoronto.ca)
• TA: Micheal Lanc (micheal.lanc@mail.utoronto.ca)
• OH: Thurs 1:15-2:15PM KW384 and by appointment
• Tutorials are primarily for discussions
  • Attendance: 50%
  • Participation: students offer high quality contributions
• Quizzes 15%
• Tutorial participation 10%
• After-class reflections 15%, you can choose to:
  • Opinion about a topic
  • Opinion about readings
  • Raise a point of confusion
• Midterm exam 20%
• Final exam 40%

Midterm

• 10 multiple choice
• 10 Fill in the blank
• 5 Vocabulary questions (what is deep learning, what is intetionality)
• 3 short answers
  • (Describe one reason for denying LLM might have genuine representation of things in external world)
  • Explain heddens argument that statistical fairness criteria besides CWG arent genuine requirements

Concepts

Week 1

• Machine Learning
• Deep Learning
• AI Scaling Laws
• Symbolic AI
• Eniac
• Turing Test
• GOFAI
• Perceptron Model
• ELIZA
• Deep Blue
• PGM
• AlphaGo Zero
• AlphaFold
• Transformer
• Generative Pretrained Transformer
• Narrow AI
• Artificial Intelligence
• AGI
• TAI
• Artificial Superintelligence
• AI Timeline
• AI Metrics
• Baptist
• Bootlegger
• Prometheus Myth
• Millenarianism
• Lump of Labour Fallacy
• Marginal Revenue Productivity Theory of Wages

Week 2

• Deep Learning
• CNN
• LSTM
• Activation Function
• Loss Function
• AlexNet
• Transformer
• Stochastic Gradient Descent
• AI Agent
• Multi Modal Agent
• AI Scaling Laws

Week 3

• LLM Memory Models
• Catastrophic Forgetting
• LLM Limitations
• Phenomenal Consciousness
• Intentionality
• Alan Turing
• Church Turing Hypothesis
• John Searle
• AI Semantic Gap

Week 4

• Does ChatGPT Have a Mind
• Interpretationalism
• Representationalism
• Blockhead
• Chinese Room
• Minds vs Simulated Minds
• Network Increase Effect
• Stochastic Parrot
• Occurant States
• Dispositional States
• Probe
• Semi Perpendicular Text Embeddings
• Intentionality
  • Information Criteria for Intentionality
  • Causal Power Criteria for Intentionality
  • Folk Psychology Criteria for Intentionality
  • Structural Isomorphism Criteria for Intentionality
  • Teleosemantics Criteria for Intentionality
• Golden Gate Claude
• Othello GPT
• LLM Stochastic Parrot Argument
• LLM Memorization Argument
• LLM Sensory Grounding Argument
• Moral Patiency
• Emergent World Representation

Week 5

• Fairness
• Differential Privacy
• Pareto Curve
• Pareto Distribution
• Hume Internal Representation
• Social and Ethical Issues in AI
• AI Fairness Issues
• AI Bias

Week 6

• Calibration Within Groups
• Brian Hedden’s Fairness Argument