Listen to the audio version of the article
For several decades now, the scientific community has focused its attention on the growing problem of people’s loneliness. In fact, loneliness is a growing phenomenon globally and the situation has worsened further following the COVID-19 pandemic. This pervasive feeling not only has an emotional impact, but also a physical one, on the sufferer. According to the American Association of Retired Persons (AARP), it is estimated that one in three people over 45 in the United States suffers from chronic loneliness. Even in the United Kingdom, loneliness has such an important social impact that a parliamentary commission has been set up to deal with the problem on the proposal of MP Jo Cox.
The question that should therefore be asked is: can artificial intelligence and conversational systems contribute positively to limiting loneliness and the negative effects it has on people’s health?
The seemingly trivial question does not, however, suggest an obvious answer. Already in 2018 the magazine of the Massachusetts Institute of Technology – MIT Sloan came out with an article entitled “Even If AI Can Cure Loneliness — Should It?” (Even if artificial intelligence could cure loneliness, should it do it?) in which intelligent systems were recognized as having the potential ability to create comparison and companionship in people, but which contextually contested the possibility that this shortcut in creating and cultivating relationships could somehow upset social norms and limit the birth of genuine relationships between people.
Indeed, the fact that artificial intelligence tends to learn from people’s behavior and adapt its responses to what the end user considers more preferable would mean that the intelligent agent tends to be a system that the user perceives as free from defects and which tolerates – conversely – any defect of the user, polarizing their behavior and beliefs (which, on the other hand, in a normal relational dynamic, tend to find natural moderation).
As often happens, a possible dystopian future has been anticipated by the cinematography and in particular by the 2013 film “HER”, in which the plot revolves around a man who – abandoned by his partner – develops a sentimental relationship with an operating system endowed with intelligence artificial and in which – in a significant scene – a woman, played by actress Scarlett Johanson, lends herself to giving body to the operating system itself to allow for a physical relationship with the protagonist (the man therefore – in an inversion of roles – becomes a tool of the machine).
This limit situation, proposed in the imagination, has materialized – after almost ten years – only recently with the case of Replika. Replika is, in fact, an app designed to interact with users and become their friend. Following an update, the software has partially changed its behavior and some loyal users have experienced this update as the loss of a loved one, even threatening self-harm.
This situation must be an alarm bell, a warning that suggests that, although artificial intelligence and chatbots in particular offer promising solutions to address the growing problem of loneliness, it is essential to reflect on the ethical and moral implications of these solutions and that they to support and not to replace genuine and healthy traditional social interactions.
* Research assistant at the University of Pavia and professor of Digital Marketing – Rome “La Sapienza”