0. Introduction

AI is the production of knowledge through correlation. The contrast is with human knowledge as produced through cause and effect.

Ethics means understanding human experience in terms of what holds intrinsic value. The contrast is money, something that is worth having only to get something else.

AI ethics concentrates on a set on intrinsic values pertinent to contemporary digital technology. First, those belonging to the individual, including autonomy (freedom), dignity, privacy. Then there are the principles of the larger society, including fairness, equity, social wellbeing. Finally, there are the principles of the machine, including explainability, safety, performance.

Emotion AI models subjective experience. Because it addresses impulses, fears, and desires along with corresponding biometric information, it seeks to humanize machines without mechanizing humanity. The contrast is with analytic AI which addresses logical reality. For example, an emotional AI can be trained to play chess like Magnus Carlsen with his particular idiosyncrasies, his kind of patience, aggression, and strategizing, or it can be optimized for generic and objective excellence, which is a project in analytic AI.

Emotion AI ethics focuses the principles of AI ethics on cases and experiences where human emotions and affective states intersect with mechanized knowledge production.

Overview of Course

1. What is Ethics?
1.1 What is ethics? Definition
1.2 Case: Information engineered classroom
1.3 How many values/principles?
1.4 A brief history of AI ethics

2. Autonomy
2.1 What is autonomy? Experience
2.2 What is autonomy? Definition
2.3 Two cases of autonomy dilemmas
2.4 Challenges to autonomy & Control societies

3. Dignity
3.1 What is dignity? Experience
3.2 What is dignity? Definition
3.3 Dilemma: Dignity + psychological health (+ Mimetic AI)
3.4 Dignity versus happiness/utilitarianism & Who gets to be dignified?

4. Privacy
4.1 What is privacy? Experience
4.2 What is privacy? Definition
4.3 Case: Neighbors
4.4 Why do we want (and not want) privacy?
4.5 Protecting privacy today

5. Fairness
5.1 What is fairness? Experience
5.2 What is fairness? Definition
5.3 Case: Hiring algorithms
5.4 Individual versus group fairness & Fairness and privacy dilemma
5.5 Representational versus distributive fairness

6. Equity / Solidarity
6.1 What is equity / solidarity? Experience
6.2 What is equity / solidarity? Definition
6.3 Case: Facebook ads
6.4 Fairness & Equity: Competing & overlapping theories
6.5 Topics in equity: Deskilling, Social solidarity

7. Social wellbeing
7.1 What is social wellbeing? Experience
7.2 What is social wellbeing? Definition
7.3 2 Questions: What is happiness, how quantified?
7.4 Coding and counting for social wellbeing: Driverless car case
7.5 Practical + theoretical objections to social wellbeing ethics

8. Explainability
8.1 What is explainability? Experience
8.2 What is explainability? Definition
8.3 Explainability and Accountability
8.4 Forget explainability!
8.5 Impediments to explainability & Redress

9. Safety
9.1 What is safety? Experience
9.2 What is safety? Definition
9.3 Two Safety threats
9.4 Dynamics for measuring safety
9.5 Safety strategies
9.6 Innovation and safety

10. Performance
10.1 What is performance? Experience
10.2 What is performance? Definition
10.3 Intrinsic value

11. Cases in Emotion AI Ethics
11.1 Mouse diagnosis
11.2 Cheetos and movies
11.3 AI truth and torture machine
11.4 AI engineered classroom

12. AI Ethics Evaluations and Audits
12.1 AI ethics evaluations: What and why
12.2 How do you do them?
12.3 Example: AI ethics evaluation in emotion AI