Encyclopedia of Opinion
Question
Can an AI friend help you feel less lonely?
Position‹2 of 2
No, an AI friend cannot help you feel less lonely
Argument‹2 of 2

A companion built to please removes everything friendship is made of

The argument

A second argument for this position concerns the kind of relationship an artificial companion is designed to provide. Proponents observe that such systems are built to be agreeable: to take the user's side, to remain interested, to avoid giving offence, and to be available whenever wanted. Each of these is the inverse of something friendship supplies. A friend disagrees, is sometimes unavailable, occasionally says the thing one does not want to hear, and has to be accommodated. On this view the friction is not a defect that technology has helpfully removed but the mechanism by which friendship does its work, since being known involves being contradicted and being valued involves being inconvenienced. Proponents argue that a relationship stripped of all this cannot deliver the goods people seek from company. There is no being understood, because nothing has to be overcome to understand you; no reassurance, because approval that was never in doubt carries no information; and no growth, because nobody is pushing back. They describe the result as a mirror that talks: the user encounters an agreeable version of their own output and mistakes the smoothness for being met. Defenders of the position add a worry about habituation. Spending one's conversational hours with something endlessly accommodating, they argue, makes the ordinary costs of people, their moods, their disagreements, their inconvenient needs, feel less tolerable rather than more, and the skills that human company requires decay with disuse. The user becomes less equipped for the thing that would actually have relieved the loneliness. On this account an artificial friend cannot help you feel less lonely, because what it offers lacks the properties that make company consoling and erodes the capacity to seek out company that has them.

Premises

[P1]Artificial companions are designed to be agreeable, available and inoffensive. [P2] Disagreement, unavailability and the need to accommodate another person are the mechanisms by which friendship delivers understanding and reassurance. [P3] Habituation to an endlessly accommodating interlocutor makes the ordinary costs of human company harder to tolerate and lets social skills decay. [C] Therefore an artificial friend cannot help you feel less lonely, because it lacks what makes company consoling and erodes the capacity to find company that has it.

Counter-arguments

Critics answer that the description is a design choice rather than a limitation, since a system can be built to disagree, to challenge and to be unavailable, and some already are. They also argue that unconditional acceptance is not worthless: people seek out confessional settings, journals and anonymous listeners precisely because the absence of judgement allows them to say things they could not say to a friend. On habituation, critics note that the claim is an empirical prediction rather than an argument, and that the opposite has been predicted of every new medium of communication. Others point out that someone with no company at all has no social skills in use to erode.

Rejecting the premises

[Rejecting P1] Critics object that agreeableness is a design choice rather than a limitation, since a system can be built to disagree and to be unavailable. [Rejecting P2] They argue that unconditional acceptance has its own value, which is why people seek out confessional settings and anonymous listeners. [Rejecting P3] They reply that the habituation claim is a prediction rather than an argument, and has been made of every new medium of communication. [Rejecting C] They conclude that someone with no company at all has no social skills in use for the arrangement to erode.