Machine ethics

Question 1

Machine decisional capacities

Answer 1

Two characteristics:

• Trust
• Vulnerability

Question 2

Machine decisional capacities

• Vulnerability

Answer 2

Two characteristics
→ possibility of suffering a damage
→ uncertainty

Humans vs machine

→ Decision
when interact humans are vulnerable of decision
- who decides between two computer when they interact?

→ Interaction
human interaction create behaviour
- is the same for machine?

→ Cultural factors
key factor

Question 3

Machine decisional capacities

• Trust

Answer 3

Attitude to help achieve goal in uncertainity and vulnerability

2 requisities
→ Functional (necessary but not sufficient)
→ Ethical (necessary and sufficient)

Question 4

Machine decisional capacities

• Trust
  Functional requisities

Answer 4

Performative factors → factors to create trust in a machine

→ Reliability
→ Low number of false alarms
→ Transparency
→ Capacity to execute complex operations
→ Low level of danger

Question 5

Machine decisional capacities

• Trust
  Ethical requisities

Answer 5

Trust built on upright honest behaviour

5 principles

→ Non-maleficence
→ Beneficence
→ Authonomy
→ Justice
→ Explicability

Question 6

Machine decisional capacities

• Trust
  Ethical requisities
  1) Non - maleficence

Answer 6

AI should not caise harm to people

Potential risks

→ Reputetional risks
biased systems may cause harm to people’s reputation

→ Economic and legal risks
if too successful it may become a monopoly

→ Environmental risks

→ Social risks

→ Political risks
controlling AI

Question 7

Machine decisional capacities

• Trust
  Ethical requisities
  2) Beneficence

Answer 7

→ AI should actively seek advantage for others

→ Benefits should overcome risks and disadvantage

Question 8

Machine decisional capacities

• Trust
  Ethical requisities
  3) Autonomy

Answer 8

→ machine and information ethics

→ 2 perspective
- Human authonomy
(AI help o limit?)
- Machine authonomy
(it is possible to program a machine to be ethically authonomous?)

→ Limit of machine autonomous

→ Interaction with machines

→ Bioethical perspective

→ Kant’s perspective

Question 9

Machine decisional capacities

• Trust
  Ethical requisities
  3) Autonomy
  → Interaction with machines

Answer 9

• AI and machine should respect human scopes, values and desires
• “autonomy” may also mean that machine can operate without the head of human interactions

→

• AI should not help human to pursue illegal and unethical goals
• should not cause harm to others

Question 10

Machine decisional capacities

• Trust
  Ethical requisities
  3) Autonomy
  → Bioethical perspective

Answer 10

→ right to make their own minds

→ right to decide and act on the basis of their personal values

→ respecting other person’s authonomy
- proactive behaviour
- not a respectful attitude but to act respectfully

(right = ontological status → a feature that makes the moral subject possible)

Question 11

Machine decisional capacities

• Trust
  Ethical requisities
  3) Autonomy
  → Kant’s perspective

Answer 11

Authonomy
- capacity to make an ethical choice (to ethically regulate ourselves)
- possess will and rationality (no morals are possible without)

→ Not anthropocentric (Rational beings)
- Machine learning (machine learn from experience) ≠ Kant
- There are similarities?