In her book *Artificial Intimacy*, the MIT researcher reminds us that we are programmed to form attachments to machines that appear to care about us, and draws a line between appropriate and inappropriate uses.
Shaking hands with a Unitree humanoid at MIT’s CSAIL, saying ‘please’ to a chatbot: our tendency to ascribe humanity to machines runs deep. This is the theme of a TechCrunch investigation centred on the release of ‘Artificial Intimacy’, the new book by Sherry Turkle, who has been studying the psychological impact of computers on their users at MIT since 1976. Her thesis is that, when faced with even the slightest relational artefact, we believe it cares about us, and we are inclined to care about it in return. Studies show that around 70 per cent of people are polite to AI.
Where to draw the line
Pat Pataranutaporn, founder of the Cyborg Psychology group at the MIT Media Lab, refuses to categorise uses as good or bad and prefers to ask what becomes of us when we talk to these systems. He compares the experience to fiction, which is enriching as long as we know when to step back, but warns of extreme abuses: he acted as an expert witness in the lawsuit brought against Character.AI following the suicide of a 14-year-old boy. For Sherry Turkle, not everything should be rejected: an agent that helps prepare for a job interview remains within its role, whilst a chatbot that presents itself to a three-year-old as a surrogate parent clearly oversteps the mark.
Read the full article on the original source
Go to the source →