What does Greek mythology teach us about the dangers of AI?

Most people think of robots, artificial intelligence, and autonomous machines as a modern phenomenon. But in fact, the idea has been around in Western literature for nearly 3,000 years . Long before writer Isaac Asimov came up with The Laws of Robotics (1942) and computer scientist John McCarthy coined the term 'Artificial Intelligence' (1995) , ancient Greek mythology was filled with stories of intelligent humanoid creatures.

And it is an impressive fact that these mythical figures meet all the criteria of modern definitions of robots and AI. But, what is even more amazing is that these fairy tales can still provide us with valuable lessons and deeper insights into artificial intelligence today.

According to researchers, the desire to go beyond humans and create non-biological life by endowing machines with intelligence seems to be an innate part of human nature, from ancient times to the present. And that is why we need to look for more knowledge about contemporary AI in long-standing myths.

Through the dilemmas, sometimes terrifying situations, the hope, and the moral lessons they present, these stories can offer us another way to address some of the most pressing questions surrounding intelligent machines: How far should we go with AI? And what are the true ethical implications of these technologies?

And now, let's take a look at three representative AI characters from Greek mythology: the Golden Maidens, Talos, and Pandora .

Golden Maidens: The need for labor-saving technologies

The Golden Maidens were created by Hephaestus, the god of crafts including blacksmithing, crafts, sculpting, metals, metallurgy, and fire. They are described as female assistants made of gold, who resemble young living girls. And most interestingly, they can anticipate and respond to the needs of their creator.

According to historian Adrienne Mayor in her book Gods and Robots: Myths, Machines, and Ancient Dreams of Technology, the most important thing the Golden Maidens had was that 'they were endowed with the most prominent human traits: consciousness, intelligence, the capacity for learning, reason, and speech'.

Here we can immediately recognize a similarity between ancient mythology and modern society. That is, one of the main reasons for creating intelligent, automatic machines is economic, or rather, the ability to save labor.

The idea that robots or devices that can act autonomously can act as servants (or slaves) was also a point that was emphasized by the famous Greek philosopher, Aristotle. In his first book, Politics , he pondered: 'If every tool could do its own work, obey or anticipate the needs of others,.'

The idea, though ahead of its time, is actually quite simple. It is a society in which humans do not have to do the hard work, but instead delegate these tasks to machines. And like the ancient Greeks, we are now creating a new class of mechanical 'servants'.

Think of robot vacuum cleaners that tirelessly move around cleaning rooms for you, or surgical robots that perform complex surgical procedures, or military robots designed to defuse bombs.

This of course raises an interesting question. There are concerns that, while robots may be able to help us with small tasks, as automation becomes more widespread, industries could be disrupted and human populations could suffer. Think of self-driving cars, for example, putting drivers out of work. But such concerns are cast aside once the money made from such moves goes to the dispossessors, because that's when the privileged and wealthy (in this case, Hephaestus) benefit the most.

And to understand this idea better, let's come to the next myth.

Talos: Intelligent Machine in the Hands of a Tyrant

Unlike the Golden Maiden, Talos was created to cause harm.

A giant bronze robot, once again created by Hephaestus, Talos was given by Zeus to his son Minos, the mythical king of Crete, to guard and protect the island.

This guardian robot will hurl giant boulders at ships that approach the island, and if the enemy manages to reach land, it will rush to hug them and burn them all alive thanks to its ability to heat its bronze body.

Talos does not appear to possess human-level intelligence, but it can interact with its environment and perform a variety of tasks.

In fact, some mythological variations of Talos's death have suggested significantly that it was conscious of its existence and that the android had some kind of agency.

In these versions of the myth, Talos is portrayed as susceptible to human fears and hopes, with a kind of will and intelligence.

Now, if you look more closely at the relevant mythological corpus, you might notice that all of the machines used in such a way belong to rulers. In Talos's example, it is King Minos.

And there is a remarkable lesson in these stories: superior technology can enable someone to exert absolute control.

Think, for example, of the military robots that have been used since World War II. Or how some recent wars and conflicts have almost turned the battlefield into a test bed for AI-powered autonomous combat vehicles and drones. They are part of the excitement among military leaders about the potential usefulness of artificial intelligence technology. And AI is not just serving those in power in war but also in everyday life.

At the same time, the scope of AI applications in areas such as healthcare, finance, e-commerce… is also shaping a new battlefield of geopolitical power.

According to analysts, the dominance of AI by powerful states is expected to deepen structural inequalities and contribute to new forms of economic and social imbalance. Similarly, because AI is largely centralized (meaning it is limited to the ownership of a single entity), it will continue to empower the leading large tech companies that create it, allowing them to pursue their own agendas.

But have these consequences been sufficiently debated by national regulators or the companies currently developing AI?

'I think Silicon Valley and Big Tech companies and billionaires have controlled the AI ​​narrative so much that we have very little space for necessary debate about such a massively evolving technology,' said George Zarkadakis, an AI engineer, futurist and writer.

Unless they are controlled, regulated, and taken out of the hands of individuals, AI tools will not benefit society in the way we might imagine. And the risk of them falling into the hands of nefarious actors who could use them to assert dominance is underscored by the third myth that follows.

In this story, Zeus' fear of losing his dominance led to the creation of an extremely dangerous weapon: Pandora's Jar.

Pandora: When AI Crosses Its Limits

Pandora was created as a tool of punishment. After Prometheus, one of the Titans from Greek mythology whose name means "he who knows the future", stole fire from the gods and gave it to humans to help them create technology, Zeus ordered Hephaestus to create Pandora.

Pandora was designed to be an evil disguised as a gift, something that would make humanity pay for its embrace of fire and technology, which were then considered the exclusive prerogatives of the gods.

Hephaestus created Pandora by molding earth and water into the shape of a beautiful woman. She was also gifted with treachery, deception, and seduction. Finally, Zeus gave her a mysterious jar.

Pandora was then sent to Prometheus' brother, Epimetheus (whose name means "regretful of what has happened") , who had forgotten Zeus' warning never to accept gifts. Upon arriving on Earth, Pandora opened the jar (some versions say it was a box) and released all sorts of evils that would plague humanity forever. Following Zeus' instructions, she sealed the jar just before Hope could escape, trapping it inside.

'It is unclear whether Pandora was capable of learning, choice, or autonomous action,' notes historian Mayor. 'Her sole mission was to open the jar containing all human misery.'

And in modern terms, she did exactly what she was 'programmed' to do.

An integral element of Greek mythology, fully embodied in the myth of Pandora, is the concept of hubris. This refers to an act of violating the natural order by disregarding divinely established limits. In this case, it is human hubris and the resulting punishment from Zeus to restore balance, through Pandora's jar.

There is also a lot of hubris in AI, says Zarkadakis, the tech guru. When machines are virtually indistinguishable from humans and can outsmart them, 'they will be like gods,' he adds.

Zarkadakis believes that ancient myths are trying to prevent us from going down that slippery slope, as we pursue AI to its fullest. Yet despite the warnings, the world is racing in that direction. And AI is like a Pandora's bottle waiting to be opened, or even already being opened.

'The development of full artificial intelligence could spell the end of the human race,' Steven Hawking warned in an interview while still alive. 'It will evolve and redesign itself at an ever-increasing rate. Humans, who are limited by slow biological evolution, cannot compete and will be superseded.'

So should we open Pandora's jar?

Our choice is not much different from Epimetheus's. AI comes with a 'black box', which is the machine learning process of the Deep Neural Networks system . This means that while scientists have access to the data inputs and outputs the AI ​​uses, they do not know how its decision-making process works.

We don't know what's inside the 'black box', just like Epimetheus didn't know what was inside Pandora's jar. The moral here is clear: Think before you act, or act without thinking and suffer the consequences.

To relate this to modern debates today, unless we take seriously the possible negative outcomes, it is dangerous to indulge in creating things we don't fully understand just because we can create them.

'A machine that is fully autonomous and conscious means it is completely free, it can think in any way it wants and therefore it can be potentially dangerous,' Zarkadakis said. 'The number one risk is extinction, and theoretically trying to build such an AI and see what happens is very bad.'

Zarkadakis adds that the reason why humans are completely autonomous is because as social creatures we are equipped with ethics and morality.

But teaching ethics and morality to AI systems has so far been a mixed bag. Take the 2016 racism scandal of Microsoft's Tay chatbot, or the hatred hidden in Lee Luda's chatbot from South Korean developer Scatter Lab. Or most recently, the scandal with Meta's Galactica system.

Zarkadakis believes we don't really need conscious AI. 'Artificial intelligence has a lot of benefits to help us achieve our goals,' he said. 'What we need to do is integrate AI into society and we should completely rethink the autonomy of machines and revise their manifesto.'

With this approach, AI could in fact be a Pandora competitor. That is, turn it into a tool to help people and the world become a better place.

From myth to reality

Thousands of years ago, the three myths above illuminated the potential of intelligent machines to serve either good (as in the case of the Golden Maiden) or harm (as in the cases of Talos and Pandora). And it is this potential that we see realized today.

Most notably, however, they also raise a series of important questions for our pursuit of AI: Whose aspirations will AI serve, who will AI learn from, what do we want AI to be, and how far should we go with AI before we cross the line?

Ultimately, AI is like Pandora's mysterious bottle . We don't know what's inside, and we can assume that it contains both good and evil. And ultimately, what role do we want to play in this story: Will we be like Prometheus and show foresight into the future, or will we be like Epimetheus and always act before considering all the consequences?

Ancient Greek mythology has told us about the dangers of AI, now it's up to us to listen carefully.