Medea the hacker, Pandora's jar the black box

Mrs Mayor, how did you first come to think about the connection between AI and robotics and Greek mythology?
I have been working at Stanford University in California since 2006, and am therefore surrounded by ever-advancing innovations in technology about robotics and artificial intelligence. I myself am a scholar of Greek mythology and a historian of ancient science. To me, it seemed natural to ask about the origins of such modern-seeming dreams and endeavors. I wondered: Was it possible that self-driving devices, automata, and other forms of artificial life could have been imagined in antiquity – long before technology made such machines possible?

And what did you find out?
That apparently, it was! The myths and historical documents I studied reveal that people millennia ago could not only imagine many of the concepts underlying robotics and AI, but that they also anticipated the practical and ethical problems that surround such technologies today.

You write in your book, Gods and Robots which was published in 2018, that in ancient mythology as well as history, many automata were made to harm humans. What does that tell us?
Well, basically, it suggests that the links between tyranny and technology have posed an eternal threat since antiquity. And that we should pay attention to who seeks to develop robotics, AI, and surveillance technologies and why. Advances in robotics and AI – such as facial recognition, lip-reading cameras, and robust, super-agile robots – are usually promoted to the public as beneficial. But it is easy to discern nefarious military and corporate applications and misuse by autocratic entities.

Let’s dive into some of the myths you mention in your book. For example that of Epimetheus and Prometheus ...
Yes. It is one of the oldest Greek myths: The father of the gods, Zeus, commissioned Hephaestus, god of invention, to design an artificial woman: Pandora. Her purpose was to punish humankind for accepting the gift of fire stolen from the gods by Prometheus. Zeus sent this lifelike android to earth, carrying a sealed jar filled with misery for mortals. Pandora was presented as a bride to Epimetheus, brother of Prometheus. Little did he know that Pandora was “programmed” with one mission: to open the fateful jar. Prometheus tried to warn his brother not to accept Zeus’s dangerous gift. But Epimetheus was charmed by Pandora and welcomed her into his life. Only later did he come to realize his terrible error.

How does this myth relate to current debates about robots, AI, and machine learning?
Well, for one thing, Pandora’s jar resembles a kind of black box, which once opened could never be resealed. Second, Prometheus’s name means “foresight” – he was always looking ahead, unlike his carefree brother Epimetheus, whose name means “hindsight”. Nowadays, the worried “Prometheans” among us are concerned about humanity’s future, in contrast to optimistic “Epimetheans” who are easily dazzled by short-term gains.

In Greek mythology, Prometheus is also said to be the creator of humankind. How does this myth blur the distinction between humans and automata?
The myth of Prometheus fabricating the first proto-type humans is illustrated on small gems for rings and seals: Instead of molding the forms of men and women from clay to be magically brought to life, as in the creation of Adam and Eve in the Bible, Prometheus is shown as an engineer or craftsman. He uses tools to assemble the first human being from the inside out, constructing the skeleton as the framework. This myth taps into this timeless, universal idea, that we could be soulless machines manipulated by superior beings, automata of the gods. Since antiquity and in many cultures, this sci-fi idea poses a profound philosophical puzzle about autonomy, free will, and human agency.

In your book, you also mention the first automatic garage door.
Homer describes several marvelous self-moving devices fabricated by Hephaestus for the gods. One example is a set of automated gates, which open and close on their own, to allow the gods and goddesses to drive their chariots back and forth from the Heavens to the Earth.

Another link to current technology are the Phaeacian ships. What makes them the first imagining of a GPS?
This story is part of Homer’s Odyssey: When Odysseus visits Phaeacia, a mysteriously advanced city, he tells the king of the city that he wishes nothing more than to return to his home island after ten years of wandering. The king offers a voyage on one of his fabulous ships, which sail on their own, with no captain and no rowers. Each ship is endowed with navigation maps of the known world: if you give the ship your destination, it plots the route. These ships with their access to vast data archives and charts, really seem to prefigure our advanced GPS systems.

You write that the archaic tale about the Argonauts is one of the earliest observations that cyborg or robot soldiers will bring problems of command and control. How so?
That tale mentions a giant bronze android automaton called Talos, created to defend the island of Crete. He was self-moving and could pick up and hurl boulders to sink enemy ships. His internal workings were powered by ichor, the life fluid of the immortal gods and the system was sealed by a bolt on his ankle. But the sorceress Medea was able to persuade Talos to allow Jason to remove the bolt on his ankle. When Jason removed the bolt, the robot’s power source bled out and he was destroyed. One lesson here is that automatons do not always perform as expected and they may even develop the ability to make decisions on their own, leading to disaster. Another lesson is that no matter how advanced the technology, there will always be a “hacker” like Medea to exploit vulnerabilities in the design!

A lesson is also entailed in the well-known myth of the fall of Icarus: Does it aim at cautioning against the overreaching of human limits?
It is true that the myth of Daedalus and his son Icarus is often taken as a cautionary tale against hubris. But I look at it a bit differently. After all, Daedalus successfully designed and fabricated wings for himself and his young son to escape the prison of King Minos on Crete. The wings worked – they were able to fly away. But young Icarus ignored his father’s warnings not to fly too close to the sun or too close to the sea, because heat and moisture could melt the glue in the wings. So Icarus plunged to his death because he did not follow the user’s manual. Daedalus mourned his son, but then flew on to freedom and more inventions. The moral seems to be that innovations can succeed but sometimes at great cost.

What might be the great cost with regard to AI and robot technologies?
Regarding AI and robot technologies, it sometimes seems to me that those who commission, manufacture, and deploy them seek short-term gains without considering worst case scenarios, long-term and/or unintended consequences for individuals and society, and the tendency for technology to favor tyranny. Moreover, with machine learning AI will access unimaginably vast and complex data and make non-transparent black box decisions keeping users as well as makers in the dark.

Would you say, you are opposed to advancement in these technologies?
I definitely agree with those AI techno-thinkers and philosophers who warn against reckless development of AI before we understand how to control it. And I agree with those who argue that AI must be considered as our tools, as human-made property, not as “companions,” “a fellow life-form,” or “quasi-humans” with agency or “rights”. Otherwise, who can we hold responsible for the negative actions of AI?

Another interesting parallel you draw in your book, is that between modern robots and slaves in antiquity.
Yes. Interestingly, in the fourth century BC, Aristotle described slaves as “living tools” in his justification for slavery. If only, he mused, tools and implements could anticipate our every need, like good slaves. Then, inspired by myths about automated devices and artificial beings like Talos, he wondered, what if we could have looms that could weave on their own and musical instruments that could play themselves? We would no longer need slaves, he concluded.

From today’s standpoint, it did not exactly turn out how Aristotle envisioned it ...
No, on the contrary: the old futuristic dream that AI and robots could make everyone’s life luxurious and easy seems rather ironic, as human workers verge toward becoming more like automata laboring alongside robotic technologies.

What role does imagination play when we talk about automata and robots in antiquity?
Actually, the ancient myths of Homeric times, followed by real engineering feats, reveal that imagination has always been the spirit that drives scientific advancement. More than 2,500 years ago, in the time of Homer, people were able to envision in myth how one might fabricate artificial life, automata, androids, self-navigating ships, and other self-moving devices, if only one possessed awesome or divine creativity and skills. So the ideas and concepts of robotics and AI were thinkable long before technology made them possible.

Generally, would you say that as soon as a scientific endeavor is imagined, its realization is inevitable, be it good or bad?
These mythic thought experiments were in essence the first science fiction tales. And indeed, it is often said that where sci-fi leads, technology follows. But already the ancient Greek playwright Sophocles cautioned that humans’ brilliant creativity and relentless innovation can be used for good and bad ends.

And the myths and imaginings concerning the improvement of us humans continue today ...
Yes, they do. Human Enhancement Technologies – artificial improvements to overcome frailties of the human body and amplify natural strength, sensory abilities, etc by synthetic means – were foreshadowed in ancient myths about borrowing the “super” powers of animals and gods to compensate for human vulnerability and weakness.

At the end of your book, you write that “as humans are enhanced by technology and become more like machines, robots are becoming infused with something like humanity”. Does that mean that at some point, we will all be half human, half artificial? Or are we already?
Well, we are in fact already used to glasses, hearing aids, artificial limbs and so on. But ambitions to exceed human limits via drugs, implants, exoskeletons, human-machine hybrids, gene manipulation, neurorobotics ... the quest for longevity and even immortality – they may seem beneficial but also entail practical and moral risks that we need to acknowledge as we rush headlong into the Age of Robo-Humanity.