AI Concerns and Controversies

What you will learn

• Explain the alignment problem and weigh what safety evidence does and doesn't show
• Compare the main forecasts about AI's future and name the assumptions that separate them
• Assess claims about AI's effects on work, learning, information, fairness, and creative rights
• Compare approaches to governing AI and state the strongest case for and against each

Contents

1. Module 1 Safety and Alignment 3 lessons

   This module covers the argument that advanced AI could be hard to control, and the evidence offered for and against it. You'll be able to explain the alignment problem, judge what laboratory safety tests do and don't show, and compare forecasts that run from catastrophe to abundance to ordinary technological change. The module gives you the terms of a debate in which well-informed people disagree sharply.

   Module overview and further reading

   1. 1.1Safety and Alignment: The alignment problem1h 3m
   2. 1.2Safety and Alignment: What safety evaluations and interpretability show1h 3m
   3. 1.3Safety and Alignment: Forecasts in conflict: catastrophe, abundance, and normal technology1h 33m
2. Module 2 Work, Education, and Economic Power 3 lessons

   This module covers AI's effects on jobs, on schools and universities, and on who controls the resources the technology depends on. You'll be able to weigh claims about job loss against the employment data, compare positions on AI in education, and explain the debates over energy, chips, and market power. These are the arguments most likely to touch your own work and community.

   Module overview and further reading

   1. 2.1Work and Power: Jobs, tasks, and productivity1h 3m
   2. 2.2Work and Power: AI in schools and universities1h 3m
   3. 2.3Work and Power: Chips, data centers, energy, and concentration of power1h 33m
3. Module 3 Information, Fairness, and Creative Rights 3 lessons

   This module covers three disputes about what AI does to shared goods: reliable information, fair treatment, and creative work. You'll be able to assess claims about AI-generated misinformation against the evidence, explain why fairness has competing definitions that can't all be met, and compare creators' and developers' positions on training data using what courts have ruled. Each topic has more evidence behind it than the headlines suggest, and less agreement.

   Module overview and further reading

   1. 3.1Information and Rights: Synthetic media, misinformation, and trust1h 3m
   2. 3.2Information and Rights: Bias, fairness, and automated decisions1h 3m
   3. 3.3Information and Rights: Copyright, training data, and creative work1h 33m
4. Module 4 Governance and Regulation 3 lessons

   This module covers who is trying to govern AI and how: governments with different models, companies with voluntary commitments, and international bodies with declarations and reports. You'll be able to compare the main regulatory approaches, judge what company commitments are worth and what their critics say, and explain why international agreement is hard. The module describes and compares; it doesn't say what the rules should be.

   Module overview and further reading

   1. 4.1Governance: Regulatory models: the European Union, the United States, and China1h 3m
   2. 4.2Governance: Company self-governance and its critics1h 3m
   3. 4.3Governance: International coordination, security, and military use1h 33m

Course project