Social impacts of AI systems

Part V, AI 100, 2026W1

Learning outcomes

Students will be able to:

  1. (Economic impacts) Analyze automation-related job displacement claims using the lump of labour fallacy.
  2. (Economic impacts) Predict which job categories are more/less automatable and justify reasoning.
  3. (Economic impacts) Describe scenarios for a post-scarcity economy.
  4. (Bias) Differentiate bias sources: training data, model alignment, and designer demographics.
  5. (Bias) Explain how transparency and standardization can mitigate bias.
  6. (Content creators) Evaluate the legal and cultural implications of generative AI on fair use and creative craft.
  7. (Education) Assess AI’s impact on student assessment and curriculum design.
  8. (Education) Propose strategies for adapting teaching in AI-rich environments.
  9. (Human relationships) Analyze the developmental and societal effects of AI companions.
  10. (Human relationships) Compare AI-based relationships to those that emerged from prior communication technologies.
  11. (Military use) Describe AI-enabled military technologies (e.g., drones, missile defense).
  12. (Military use) Debate the security and geopolitical implications of military AI.
  13. (Climate) Assess the environmental costs of AI training and inference.
  14. (Climate) Identify AI applications that could improve energy efficiency and reduce emissions.

Theme: Examine the broad economic, ethical, cultural, and environmental implications of AI’s growing role in society.

Select topics

Original category 1 proposal topics

  • Economic impacts
    • Lump of labour fallacy
    • Jobs that are likely to be automated
    • Jobs that are likely not to be automated
    • Jobs that are likely to be created
    • What would a post-scarcity economy look like?
  • Bias
    • Bias introduced via training data
    • Bias introduced via alignment
    • Bias introduced via AI designers being poorly representative of broader society
    • Bias reduced via standardization and transparency (everyone gets the same AI system; the system can be audited)
  • Content creators
    • Relationship to fair use
    • The role of taste vs the role of craft
  • Education
    • How is assessment affected?
    • In what ways should we change what we teach?
    • How can the role of the teacher change?
    • What skills do students need to develop to learn in this new environment?
  • Human Relationships
    • Social impacts of AI companions, AI therapists
    • Developmental impacts of these tools being used by children
    • Downstream social consequences; contrast with social networking
  • Military use
    • Current state of the art; lessons from recent military conflicts (drone warfare; missile defense; surveillance)
    • What’s around the corner?
    • Will these changes make the world safer or less safe?
    • What political structures will they foster (insurgency vs large states; democracy vs authoritarianism)
  • Intentionally harmful use
    • Deep fakes
    • Disinformation
    • Risks to privacy, safety, security
  • Climate
    • Energy intensity of generative AI
    • Prospect of reducing carbon emissions
      • Making other systems more energy efficient
      • Advancing electrification of transportation
      • Compatibility of AI data centers with green energy

Suggested revised topics

  • Economic impact
  • Possible biases
  • Climate impact
  • Impact on human relationships
  • Impact on relationships to governments
  • Impact on relationships to companies (surveillance capitalism)
  • Inadvertent vs. intentional negative impacts
  • Human-machine relationships

Dongwook’s suggested topics

  • Work: work, productivity and the economy
  • Culture: culture, creativity and education
  • Fairness: bias, fairness, power, and inequality
  • Power: governance, law, and poliitics
  • Sustainability: sustainability, infrastructure, and global impacts
  • Futures: future trajectories, existential risk, and philosophical questions