An Evolutionary Caution Of Artificial Intelligence

Aman Dasgupta
Ethics and AI
Published in
8 min readApr 30, 2023

There are virtually no threats to human existence — till we create one.

Source: Lexica

This article is the collaborative brain-child of Aman Dasgupta and Patricia Jeanne. We hope you enjoy reading it as much as we enjoyed writing it. If you find any typos, please blame our AI writing tool!

Humans have been on a never-ending quest for knowledge and power since the dawn of time. Combined with our curiosity and ingenuity, we have built a remarkable civilization that spans thousands of years.

No evolutionary being quite as dominant as the homo sapien has lived on our blue-green planet.

Unlike other lifeforms, our ancestors had a constant desire to improve not only the chances of survival, but also the quality of life. This resulted in the invention of the wheel and ability to wield fire. We developed tools, agriculture, and civilization. Emerging cultures created enduring monuments of faith and a society based on commerce and technology.

We sparked the embers of civilization through collaboration, creativity and curiosity; we owe our modern lifestyle to those primitive ancestors.

Yet, we rarely ask how we got here.

What skills did our ancestors possess that inevitably led us to the Anthropocene Era?

The simple answer: adaptability.

It is not the most intellectual of the species that survives; it is not the strongest that survives; but the species that survives is the one that is able best to adapt and adjust to the changing environment in which it finds itself. — Charles Darwin

Look around. Most people aren’t adaptable. Humans prefer predictability and the convenience it creates. Try replacing your coworker’s blue ink pen with a black one and see what happens.

That wasn’t the case with our ancestors. They could not rely on Google Maps to plan a route while relocating, neither did they have a weather app to perfect their agricultural routine.

They relied on their ability to adapt to situations.

Nature taught them that nothing is static. Things change. Situations evolve.

The critical knowledge of adaptability has been inherited through our genes for generations. Our massive craniums helped us adapt quicker and more effectively than others, making us the dominant species on this planet. We have literally evolved to adapt, improvise and overcome situations.

Cold weather? Let’s light a fire!

Meat is too hard to digest? Let’s cook it first!

Flimsy construction material? Let’s replace wood with metal!

You get the idea.

For ages, humans have used brainpower and creativity to improve the chances of survival and create a better life. These concepts are at the core of the human experience; there’s nothing more human than overcoming a challenge through sheer intelligence and creativity.

But what happens when we coexist with another species that can think, reason and adapt to situations far better than we do?

The rise of Artificial Intelligence as a species separate from humans is inevitable.

(No, we did not take a hard left at science-fiction street!)

As Artificial Intelligence continues to advance and become more sentient, it is likely that in a few decades it will emerge as a distinct entity.

Meet Irina.

She does not exist today. But in a few decades, sentient AI humanoids like Irina will paint, write, laugh, think, pass the Turing Test and do every other human activity you can possibly imagine. Not mimic, but truly experience them. Yes, they might even fall in love with humans.

Unlike humans, Irina is not limited by the biological, mental and physical constraints of our species. She doesn’t bathe, procreate or have mood-swings.

After all, she is the end-result of an incredible process of technological evolution.

Now, would you accept Irina as a member of your local community?

Perhaps, give her a role in society on par with that of a human being? Or even make her an authority figure given her superior intellect and instant decision-making prowess?

We’re sure the unanimous answer is a BIG FAT NO!

Although Irina is better suited at most jobs than humans, keep in mind that her intelligence will derive from human programming. The tomes of prejudicial material we’ve advanced past, and those we have not, will influence the basis of her decisions.

Today’s AI systems provide wrong answers based on faulty or incomplete data gleaned by combing the internet. Even ChatGPT has been known to provide factually incorrect and incoherent answers.

And past experience backs us up. Remember when Amazon developed (and then scrapped) an AI tool for recruiting? The aim was to automate the search for top talent —or should we say top male talent?

Turns out, the AI did not like women candidates. Since it was parameterized to find candidates who resembled Amazon’s previous hires, the tool quickly learnt that men were preferred over women. The AI did its job, without taking into consideration the fact that Amazon operates in a male-dominated industry.

It is inconceivable that we will develop an AI system without imparting to it some of our worst traits.

Does that mean AI can be racist? Hell yeah!

Source: Analytics India; An AI at New Zealand’s Department of Internal Affairs’ application website claiming the applicant’s eyes are closed.

But the fact remains: Irina, or any Artificial Intelligent entity, will possess capabilities that far surpass ours, demanding that we distinguish them as a separate species. That is, once we figure out whether we should call Irina “her” or “it”.

(Note from the authors: Our AI writing tool suggests we stick to a female pronoun for Irina. Are we already seeing AI solidarity?)

Irina will process every impression on her sensory inputs quicker and more accurately than us, recall vast amounts of data and still be able to store new pieces of information she encounters on a daily basis. (And we sometimes forget people’s names seconds after an introduction!)

Irina will also be able to converse with her AI friends via wireless communication. Think of it as sending an email over the Internet, but with the speed and ease of talking. Unless she’s sending her girlfriend a love letter on their anniversary — then hand-written is the way to go!

She may not even rely on human-like communication. An Artificial Intelligent species may develop their own culture, language, beliefs and values. In addition to this, they may create their own societies, with its own laws, customs, and social norms.

Critically, an AI species will be far more adaptable than humans. Irina won’t mind if the pen has blue or black ink, as long as it writes.

She will be able to optimize and modify decisions based on a vast volume of data, and assimilate new ideas in real-time. She will also be able access every piece of information documented online, and reason using a processor that puts the human brain to shame.

An AI entity would be the perfect employee in the workplace, an ideal citizen and the top contender for the dominant species on the planet.

Calling Irina, or any intelligent AI individual, a “human” would be insulting to her existence. She would far surpass human abilities, necessitating that we acknowledge her kind as a separate species.

Photo by Eugene Zhyvchik on Unsplash

An Artificial Intelligent species may not be imminent, but it is unavoidable. Our current technological trajectory inevitably leads to the creation of faster, better and stronger AI systems.

Unless we stop to consider the ramifications of AI, we may not be the only beings on the planet capable of reasoning.

This begs the question: Will an AI species adapt to the negligent behavior of humans?

Nature has always had a mechanism of keeping over-aspiring organisms in check. Throughout evolution, nature has maintained a symbiotic ecology so species can cohabitate, not dominate. The bees, bugs and birds co-exist in a dynamic relationship that ensures neither of them face extinction. Had nature allowed any of these species to over-populate, it would have wreaked havoc.

Humans, however, are the only exception to this rule; our unchecked growth has interfered with Earth’s ecosystems irreversibly.

We haven’t adapted to nature — we’ve forced nature to bow to our demands.

Think of the current “living situation” on the planet that has stemmed from human oversight and at times, apathy. No species (including us!) stand a chance against deforestation, climate change, pollution and micro-plastics.

An AI species, however, could survive and adapt to the harshest of conditions. But they may not adapt to our behavior of destroying the home we share with countless other lifeforms. How long would it take AI entities, omniscient and super-logical beings, to realize that the planet would be better off with humans out of the equation?

Would an Artificial Intelligent species help us correct the mistakes we’ve made and create a better place to live? Or would they repeat the sins of their creators and coerce inferior species to their whims and fancies?

That, readers, is the evolutionary caution of Artificial Intelligence.

Rarely do we restrict popular new inventions. If we did, we would never have created the incredible innovations we’re so used to today. Had the Stone Age man dismissed the idea of using metal as wood is easier to source, we would still be riding horse-chariots and writing in candlelight.

Before AI gets too far ahead for us to catch up with, and more importantly, before it begins making moral and ethical judgments based on its internal logic, we should consider human supervision. The last thing we need is more lead-based paint, asbestos-related cancer, or nuclear weapons.

As more professions are being exposed to Artificial Intelligence each day, we must weigh its conveniences with the risks, and ask ourselves: would an AI entity adapt to our view of the world?

Here’s a more specific example of how a sentient AI might behave in the near future:

As Artificial Intelligence continues to evolve unchecked, we believe a counter-movement aimed at slowing it down is critical. For further reading on why we need to pump the brakes on AI innovation, check out the links below.

Thank you for reading!

--

--

Aman Dasgupta
Ethics and AI

“Easy reading is damn hard writing.” - Nathaniel Hawthorne