Do Robots Dream?

Published 15 Feb 2021

By Aileen Wang

Since the advent of the technological age, as constant scientific breakthrough became the norm, the unknown future of a world changing more rapidly than ever more has captured the minds of countless artists, poets and writers. The future could no longer reliably be considered the mere extension of the past. Old knowledge was no longer secure in its vault. Anything could change at any moment. Could humankind journey to the centre of the earth? The bottom of the sea? The surface of the moon and the outer rim of space?

From the tentative beginnings of sci-fi under the speculative pen of Jules Verne to the sophisticated social commentary of Ursula Le Guin, utopic, dystopic, realist and fantastical visions of a technologically driven future appear before us as diverse as the fragments of light of a kaleidoscope. What are we to think of the inhabitants of Karhide, with their monthly kemmer, their ice and the familiarly human emotions that govern them? Can we really expect an uprising of mistreated, preternaturally intelligent apes?

However, simultaneously with an inexhaustible curiosity into the vast unknown of alien life, developed an equally inquisitive interest in the final outcomes of beings of human construction. Intimately knowable, created beings over whom we hold uncertain and self-conscious dominion, the future evolution of robots, cyborgs, androids and artificial intelligence has delighted and repulsed many over the years. Some preach fearfully of a robot takeover, others predict a computer singularity, still more dream of the future cyborg master race. The endless fracturing possibilities of the as-of-yet achieved future led to a blossoming of creative endeavour.

One of the key questions that tormented many was the future of a humanity in competition with AIs that relentlessly approached a standard of perfection that humans could not hope to compete with. What would happen to the redundant human race? What would become of the humans, outpaced in every way by the intelligent life they themselves had worked so hard to create and improve?

Some have chosen to take this scenario and turn it into a moral of aggrandizing human arrogance – the master, turned on by the servant who had long become their equal, even their better. Others, no doubt in reflection on the upheaval brought about in the present and past by technological achievement, postulate a world where the benefits go to only a few, and the majority languish under a robotic yoke. During the Industrial Revolution, after all, how many were driven into destitution as improved technologies rendered them skill-less, brought them seeking their livelihood in cramped and dirty cities? Today, how many people sit afraid that eventually their work will be automated, and they themselves, made jobless?

Throughout this, humans have clung to the hope that robots are unable and incapable of human empathy. A robot, we say, can never be kind like a human can. The fundamentals of human nature are not programmable. No matter how ‘smart’ an artificial intelligence is, it can never learn the human soul. We distinguish ourselves from cold, soulless robots through that indistinct, undefinable humanity that remains common and unique to us and us alone.

But as artificial intelligences such as Alexa and Siri become closer and closer to replications of humanity, this presumption seems…well, presumptuous. The goal of artificial intelligence, is, after all, to become infinitesimally close to the human mind, and eventually to overtake it. What’s to say one day we won’t have a robot whose only difference from a human is the material it is made of? The Turing Test for a machine requires that the machine speaks flawlessly the language of humanity in all senses of the term. When that day comes, has the machine replicated human nature, or has it learnt it?

If something acts and speaks and reacts in a certain way, we come to the conclusion that they must feel a certain way as well. If a machine replicates flawlessly and continuously the irrationalities, vagaries, and complexities of human emotions, what other than our own hubris can lead us to the conclusion that they, too, have learnt to feel human?

Of course, we’re far from that day still (or are we?). Before we start to worry about the feelings of machines, it’s best to ground ourselves in reality. Speculative fiction often vastly overestimates the future. ‘Back to the future’ predicted flying buses and cars, but even now that remains an unachieved dream. The possibilities of technology often don’t match up to their present capabilities, and certainly never reveal much about their present problems. But who knows? Before we meet a race of aliens, we might succeed in creating one.

Tags: Data Science Machine Learning, Deep Learning and Neural Networks Applications of Data Science