Kevin Corti, PhD candidate at the Department of Social Psychology, discusses how android science has developed since Alan Turing’s foundational work.
In the early stages of World War II, as the Allies escalated the intensity of their military campaign against the Axis powers, the mathematician Alan Turing and a British cryptanalyst team at Bletchley Park waged a secret battle against the German Navy’s Enigma cipher machine which was being used to transmit coded messages.
An extremely sophisticated device, many considered Enigma unbreakable. Believing that throwing purely human brainpower at Enigma would likely lead nowhere, Turing designed a machine known as a “bombe” that could simultaneously replicate the rotor movements of three dozen Enigma machines and discover the daily rotor settings being used by German Enigma operators. The mysteries of Enigma were uncovered, in other words, by creating a machine that simulated its functioning.
Today, Turing’s basic insight is being used within the field of android science to decode a machine far more enigmatic than Enigma: the human mind.
Android science is a field composed of researchers from robotics and the social sciences who seek to understand human psychological phenomena by designing robots (androids) that look, behave, and communicate like real humans. The basic premise goes like this: by comparing how people interact with and perceive interlocutors whose appearance and behavior is nearly, but not quite, human, to how people interact with and perceive actual humans, we can better understand how the microdetails of “humanness” affect our psychology. Statements that are illconstructed or slightly outofcontext, eyes that lack a glossy wetness, arm movements that are not smooth and fluid, lips that move slightly out of sync with the words they are speaking, and so on, can profoundly alter our perception of an interlocutor as well as our willingness and ability to engage with that interlocutor in communication.
Turing, of course, is also famous for his contribution to our thinking about artificial intelligence and philosophy of mind. He proposed that one method for judging whether or not a machine could display humanlike intelligence would be to test whether through text communication the machine could reliably convince interrogators that it was in fact a real person. Turing called this the “imitation game,” but today it is more commonly referred to as the “Turing Test.” Contemporary androids and other robots, however, are capable of much more than text output they can communicate multimodally. For this reason, android scientists evaluate these machines against what is known as a Total Turing Test, which takes into consideration the verbal (linguistic) and nonverbal (sensorimotor, appearance, emotional) communicative abilities of androids.
Androids can also be teleoperated controlled by a remote user who directs the machine’s movement and speech. Developers are currently testing whether or not teleoperated androids can enhance social communication by allowing, for example, people who would otherwise be unable to attend a meeting do so through an android surrogate. It is thought that androids with teleoperational capabilities might one day supplant more distal technologies such as video/voice chat.
Android science is still in its infancy, and many android technologies have yet to be tested beyond controlled laboratory environments. While developers of these technologies are quite bullish on the idea of androids soon being ubiquitous communication tools throughout society, it remains to be seen whether or not people would be comfortable socializing with machines that appear eerily similar to real people. One thing is certain however: as these technologies inch ever closer toward looking like, behaving like, and sounding like us, we will learn more than we could have possibly imagined about what it means to be human.
Hanard, S. (2000). Minds, machines and Turing: The indistinguishability of indistinguishables. Journal of Logic, Language, and Information, 9, 425445.
MacDorman, K.F., & Ishiguro, H. (2006). The uncanny advantage of using androids in cognitive and social science research. Interaction Studies, 7, 297337.
This blog first appeared on the website of IPOC, a research group within the DSP.