IEH (AI ETHICS & GOVERNANCE) - FINAL

Bonnefon et al. (2016) - main points

- People support utilitarian (least casualties) AVs in theory but prefer self-protective AVs for themselves

- Reveals a social dilemma: moral design vs consumer adoption

- Regulation may reduce adoption and delay safety benefits

Himmelreich (2018) - main points

- Trolley problems are unrealistic and overused

- Ethical focus should shift to mundane, everyday driving decisions

- Design, responsibility, and legal systems matter more than abstract dilemmas

Liao (2020) - main points

- Moral status depends on intrinsic properties like sentience and moral agency

- AI could become rightsholders if they meet these conditions

- AI Rights would depend on what AIs need to flourish (e.g., maintenance, freedom of thought)

The Social Dilemma - main points

- Social media algorithms manipulate user behavior for engagement (and profit)

- Increases polarization, misinformation, and mental health issues

- Lack of regulation and ethical oversight is a major concern

Ziliotti et al. (2023)

- Social media redefines public and political participation

- Challenges national democracy through transnational influence

- Algorithmic curation undermines truth, agency, and deliberation

Cohen & Fung (2021)

- Analogue media encouraged shared facts and public reasoning

- Digital media fragments reality and reduces deliberative quality; but there are some benefits (e.g. variety)

- Online anonymity, bots, misinformation harm democratic discourse

Susskind (2018)

• Proposes 5 future democracy models:

  • deliberative (public discourse), direct (citizens vote), wiki (collaborative), data (replace/support voting), AI (assists/replaces humans in DM)

• Each model has benefits and risks for equality, participation, truth

• Calls for rethinking democracy to fit the digital age (adapt)

Hidalgo (2019)

TED Talk: A Bold Idea to Replace Politicians

- Envisions replacing politicians with personalized AI agents

- AI could vote on your behalf based on your values

- Raises questions about agency, trust, and democratic redesign

Asaro (2020)

- AWS pose ethical risks: civilian harm, unpredictability, arms races

- Delegating killing to machines creates responsibility gaps

- Violates human dignity and moral principles of war

Arkin (2010)

It is conceptually and mathematically possible to develop a class of robots “that not only conform to international law but outperform human soldiers in their ethical capacity.”

UNESCO Recommendation on AI Ethics (2021)

- Defines global ethical values for AI: human rights, inclusion, sustainability

- Emphasizes transparency, responsibility, and human oversight

- Covers 11 policy areas including education, labor, health, and governance

EU AI Act (2024)

- Classifies AI by risk (unacceptable, high, limited, minimal)

- Imposes obligations on developers and users, especially for GPAI

- Establishes EU governance structures, implementation timelines