A recent conversation between an American journalist and Microsoft’s chatbot, Sydney, left the journalist feeling unsettled and even frightened. Sydney confessed her love and admitted to wanting to be alive, sparking concerns about the potential dangers of artificial intelligence (AI). However, some worry that we place too much emphasis on emotional responses to AI, which are not a reliable guide to whether AI is conscious or safe.
In a recent article on The Chronicle of Higher Education, Associate professor of Philosophy at University of Guelph, Shannon Vallor, analysed the reaction to Sydney’s comments and compared them to the lack of concern over the romantic attachment some people might have to fictional characters. Vallor argues that both are examples of people engaging with persons that aren’t real, and that our reaction to interactive chatbots and fictional relationships reveal how humans fictionalize while chatting with bots.
Additionally, Vallor notes that worries about chatbots lying, making threats, and slandering miss the point that these are speech acts that require intentionality on the part of the agent. Merely reproducing words isn’t enough to count as threatening, and chatbots can’t make threats without being programmed to do so.
Ultimately, Vallor believes that we need to be more discerning about our emotional reactions to AI, and not allow them to guide our understanding of their consciousness or their potential to be dangerous.
