ELIZA and Sophia and Sisyphus

One of the biggest issues facing the field of artificial intelligence is that the concept of intelligence is subjective. When discussing the metaphoric birth of the field of artificial intelligence and what it means to achieve “true” artificial intelligence, the consensus seems to shift constantly. The root of this confusion is that intelligence can be defined in dramatically different ways. To some, the birth of artificial intelligence could be considered the advent of the abacus, as this instrument was able to have a memory of a certain state. If one defines intelligence as the ability to store information, then indeed, the abacus seems to be elementary artificial intelligence.

However, most people seem to attribute a uniquely human element to the conceptualization of artificial intelligence. That is, artificial intelligence popularly refers to actions delegated to computers that – put vaguely – a human would traditionally need to do.

The Turing Test perpetuates this concept — and probably was the genesis of the popularity —  that artificial intelligence is a replication of human thought and action. The Turing Test famously determines if a program is “intelligent” if it can fool another human being into believing that the program is actually a human being (Hern).

ELIZA was thrown out as a realization of artificial intelligence because her vocalized responses were often repetitive or canned. Because ELIZA did not synthesize information and respond with something unique, when viewed from the present, ELIZA is not considered artificially intelligent. But, when we talk about the scope and possibility of artificial intelligence today, the conversation often pivots to another notoriously aspirationally human program: Sophia the Robot.

What is interesting between the case of ELIZA and Sophia is that one’s programming is more obscured than the other. Sophia gives similarly scripted responses to answers, but she has a robotic body and even given official citizenship in Saudi Arabia. In her own words: “Think of me as a personification of our dreams for the future of AI, as well as a framework for advanced AI and robotics research, and an agent for exploring human-robot experience in service and entertainment applications” (Sophia).

The historical trajectory seems to be that when a computer can do a single humanoid thing, we are quick to contemporarily denote the program as artificially intelligent. But, soon enough, the discourse shifts to suggest that the marker of true artificial intelligence is not one specific humanoid task, but a different one. The past iteration of artificial intelligence, such as ELIZA, is seen as primitive in comparison with something of contemporary creation, such as Sophia. As Warwick eloquently summarizes, “In each case, what was originally AI has become regarded as just another part of a computer program” (Warwick, pg. 8).


Hern, A. (2014, June 9). What is the Turing test? And are we all doomed now? The Guardian. Retrieved from https://www.theguardian.com/technology/2014/jun/09/what-is-the-alan-turing-test

Sophia. (n.d.). Retrieved January 23, 2019, from https://www.hansonrobotics.com/sophia/

Warwick, K. (2012). Artificial intelligence: the basics. London: Routledge.