Week 3: Social AI and the challenge to connection

Question 1

What is social AI?

Answer 1

AI: denotes computer programmes and performing tasks commonly associated with intelligent beings

Social AI: forms of AI that are designed to behave in ways that mirror human (or animal) social behaviour.

Question 2

Can Social AI be embodied?

Answer 2

Yes: form of humanoid and animalistic robots

No: social AI can also be unembodied: computer algorithms supporting humans

Question 3

How is social AI used in the healthcare field?

Answer 3

For example against loneliness (companion robots)

Question 3

What do Fiske et al note regarding classical bioethical principles of autonomy, non-maleficence, beneficence and justice?

Answer 3

Autonomy: informed consent may be hard (dementia, intellectual disabality)

Non-maleficence: it can harm the wellbeing

Beneficence: it can reach people who otherwise wouldn’t ask for help

Justice: needs to be evenly distributed and the use should not come at the expense of human-given care

Question 3

What does Turkle think about social AI?

Answer 3

She is very skeptical about the idea that technology can provide truly meaningful forms of social connection

Question 3

What is the particular field Fiske et al focus on?

Answer 3

field of psychiatry, psychology and psychotherapy

AI-supported virtual psychotherapeutic applications are being developed.

Positive effect: lower depression and anxiety and works better than e-books

Question 3

What does Turkle mean with the ‘robotic moment’?

Answer 3

the development of a state of mind in which we became willing to grant sociable artifacts more than their due
( moment = when we accept the connection the robot seems to give)

Question 4

What is the main idea of Turkle?

Answer 4

connection with social robots is appealing to people because it gives us a sense of companionship. But, Turkle states, that companionship is not true, because it can not understand us.
Spending time with social Robots distracts us from connection with real people due to unrealistic expectations and devaluation

Question 5

How is the development of the robotic moment enabled?

Answer 5

by the ability of social robots to appeal to human vulnerabilities: our need to be recognised and not to feel alone

‘as if’ empathy

it just pushes buttons and is easier than humans because it responds directly to one’s need

Question 6

What is the main problem Turkle has with social AI?

Answer 6

That there is no authenticity, just pushing buttons is a form of manipulation.

Question 7

What does Coeckelbergh argue?

Answer 7

Robots can become human companions if they are capable of evoking human empathy

Question 8

in which two ways can empathy be understood according to Coeckelbergh?

Answer 8

1. cognition: rational understanding of the state of the other
2. feeling: an affective resonating with the other

Question 9

which form of empathy is needed in robots? (CB)

Answer 9

empathy as a feeling needs to be mirrored

Question 10

Why does CB argue for robots as a companion for humans?

Answer 10

We shouldn’t set unfair criteria:
1. it would be unfair to require that robots have a capacity for empathy because we also don’t expect this from all human companions (babies/mental disorder)

2. when it comes to domestic animals, humans consider it sufficient that people experience empathy towards them , not necessarily that empathy is experienced reciprocally (wederzijds)

therefore: robots can be recipients of human empathy

Question 11

What does CB see as an ethical problem with evoking human empathy through the mirroring of vulnerability of humans?

Answer 11

It may be deceptive (bedriegelijk)

Question 12

Why is deception not always ethically problematic according to CB?

Answer 12

deception is present in all social interactions, and he doubts if deception can apply to empathy as feeling because feelings can’t be true or untrue

Question 13

In what way does CB’s argument go against Turkle’s concern of authenticity?

Answer 13

CB states that robots are just a machine

Coeckelbergh considers empathy-as-feeling to examine the human reaction to the robot’s mirroring of vulnerability. For Turkle, it is the robot response to human vulnerabilities, which mirrors empathy-as-cognition, that is the core of the problem. Turkle criticises this ‘fake’ empathy because it lacks the empathy-as-feeling that allows people to genuinely emotionally resonate with each other (and other living beings).

Question 14

What is a central point in the works of Turkle and CB?

Answer 14

Whether it is possible to establish a genuine connection with social AI

Question 15

How does social AI affect the agency and vulnerability of human beings in the context of healthcare?

Answer 15

• can help reduce loneliness and can stimulate patients to participate in physical exercise
• scarcity of robot companions can lead to jealousy and competition
• impact on social expectations of human others
• can lighten the tasks of human healthcare professionals
• some caretasks however require people taking care of each other

Question 16

How does social AI affect the agency and vulnerability of human beings in the context of social encounters?

Answer 16

1. may reduce social vulnerabilities and increase social agency
2. can make people feel better
3. can be a safe space for people who are harmed (domestic abuse)
4. can draw attention from real meaningful human connections (Turkle)
5. people can forget their social and moral skills (Turkle)
6. people can extend their empathy-as-feeling to companion robots (CB)
7. does it belong to the human or non-human world? (Copson: connection to other people and the non-human world is essential for living out one’s humanity to the fullest, social AI is difficult to place in a binary distinction between the ‘human’ and the ‘non-human world’. )
8. raises the question what is more central to ‘humankind’: Is it the biological status of being born a homo sapiens or distinctive forms of agency that set us apart from other animals?
9. This expansion of our notion of humankind may greatly increase the agency of humankind through embracing the potential of virtual and artificial intelligent “human beings”, but also bring about new forms of vulnerability as biological and artificial vulnerabilities do not neatly align, as pointed out by Coeckelbergh.

Question 17

How does social AI affect the way we respond to the agency and vulnerability of others?

Answer 17

Turkle: negatively affect, it distracts us from real human contact

CB: robots can have their own vulnerabilities