SLIS faculty member Hamid Ekbia gave the keynote talk at the Beyond AI: Artificial Dreams International Conference held November 6-7, 2012 at the University of West Bohemia in Pilsen, Czech Republic.
The conference focused on "interdisciplinary and philosophical aspects of artificial intelligence. The aim of this year is to question deep-rooted ideas of artificial intelligence and cast critical reflection on methods standing at its foundations. Artificial Dreams epitomize our controversial quest for non-biological intelligence, and therefore we encourage prospective authors to fully exploit this controversy in their papers. Interdisciplinary dialogue between experts from engineering, natural sciences and humanities is strongly encouraged." (conference website)
The title and themes of the conference build on Dr. Ekbia’s book Artificial Dreams: The Quest for Non-Biological Intelligence (2008, Cambridge University Press.). His talk was titled “Heteronomous Humans and Autonomous Artifacts: The Paradox of AI.” (The following account is based on the abstract for the talk.)
I focus on the development of what can be called a “relational” view of intelligence and a research agenda that would be built on this view. The relational approach is often contrasted with “substantive” approaches, which consider intelligence as a real and substantive possession that groups of intelligent beings acquire through social processes. Those adopting the substantive perspective criticize the relational approach for viewing intelligence as nothing but a set of attributions by the social group: “Intelligence is in the eye of the beholder.” This is not how I understand and use “relational.” For me, a relational perspective takes intelligence not as a property of an individual entity (a human or a machine), and not simply and merely as an attribution, but as crucially dependent on the performative capabilities of other entities that are outside the individual. Understood in this fashion, “intelligence is in the act of participation;” it is in the capability to interact and relate meaningfully with relevant others. Finally, in my way of thinking, Intelligence is a mediation.
I discuss the topics of "autonomy," "affects," and "ethics" to highlight differences between the substantive and relational views. These topics are of great interest in AI and beyond. "Autonomy," for instance, is a key motif in the design of what are called "autonomous agents," "autonomous robots," and so on. I want to challenge some of the thinking behind these notions by going back to the origins of the term "autonomy" in modern times. Jean-Jacques Rousseau, for instance, considered autonomy first as “positive freedom” (people act according to their own free will but in accordance with a social contract that they themselves create and submit to) and second as social. Rousseau’s concept has given rise to a paradox that many have commented on, and in a way can be described as the key paradox of modernism: the existence of individual autonomy in a just and humane society. On this basis, perhaps we can consider the general interest in autonomy, especially in what is happening in AI, as a projection of the modern humanity’s desire for independence and freedom, coupled with despair in understanding human neediness and dependence.
At SLIS, Dr. Ekbia is the director of the Center for Research on Mediated Interaction (CROMI) and teaches several courses related to human-computer interaction.
Posted December 17, 2012