Why AI might end up destroying much of the internet and send us 150 years back in time

Disclaimer: Opinions expressed below belong solely to the author.

Whenever a new technology is born, we have to remember that is still just a tool in people’s hands and as such, it is going to be used to fulfil our basic desires.

Aside from mostly mindless entertainment and leisure (which we use to recharge ourselves), most of human activity has two driving forces: material gain and immaterial social status, based on interaction with other humans.

In essence, then, we want to secure more resources (money) to live our lives, do what we want, settle down, have a family — and we also want to be recognised as valuable, attractive members of our community.

To that end, we use all of the tools at our disposal to improve our chances at succeeding in both — and various artificial intelligence (AI)/machine-learning driven solutions are going to be a part of this tool set just like computers, smartphones — and the internet itself — are today.

Much like with every technological revolution, these new solutions are first and foremost deployed to improve business efficiency, increase our productivity and company profits.

And in that regard, any form of more or less sophisticated AI is surely, and understandably, going to see rapid adoption both among individuals and entire organisations.

If a machine can do something more quickly, accurately and cheaply, it would be dumb not to use it.

However, while thinking learning machines can make our jobs easier and help us make more money, what about the “social” aspect of being human?

By humans, for humans

A huge part of the internet is built around social interaction.

We connect with our friends and families, we make new acquaintances over the web, we then create content, gather followers on platforms like Facebook, Instagram, TikTok or YouTube — and, in many cases, even make a living out of what we publish there.

Carefully taken, polished vacation shots on Instagram draw both inspiration and envy of our friends and fans.

Video explainers teach us about a gazillion of topics, from international politics to lore of Lord of The Rings. Meanwhile, a short TikTok reel can bring out a laugh over animal shenanigans or show us how to fry an egg properly.

This mixture of useful, entertaining, mundane, inspiring or infuriating content someone else published on the web is what drives much of human interaction today.

But what if all of it was fake?

Or, at least, what if it could be fake? What if you couldn’t trust any picture, any video clip, any tip ever played on the web? What if forums like Reddit were mostly filled with bots perfectly pretending to be humans?

(Ironically, this is how Reddit started, with its founders creating hundreds of fake accounts, pretending the site was popular.)

We have a propensity towards faking things even today, with Photoshopped photos, edited videos, staged poses and locations. But even that still takes a bit of effort — effort that AI is now promising to remove.

Computers are already good enough to process enormous amounts of diverse data and present it in a casual, human-like way a’ la ChatGPT. Feed them a few photos of yourself and you can get an entire collection of images that nobody took or drew.

They can write essays, poems, news articles, create realistic photos of people who don’t exist and even compose music. That is not a maybe, that’s a fact of today. It’s only a matter of time before they are powerful enough to create complete videos or even entire movies.

Heck, there are already entirely made up influencers — curious today, but a likely plague in the future.

What if that girl you follow for makeup tips was never born and is only a product of a crafty nerd using easily accessible AI tools?

What if the Caribbean summer holiday of your friends never happened and they just decided to make themselves look superior on social media?

What if your cousin isn’t married and has no kids in real life and all of their photos are fake? What if you spent three hours chatting away with your best buddy only to learn it was merely a digital bot which has learned to perfectly mimic his way of speaking and was fed facts about your friendship?

Heck, what if that football match you watched on TV never happened? What if the president didn’t say what the clips on the internet have shown? What if that news reporter telling you about this isn’t real either?

If all things digital can be made up to the point of being indiscernible from the real thing, can we trust anything published on the internet ever again?

We’re being told the very same technology is being used to detect if something was made by another machine — but given time, it’s pretty clear that it’s just delaying the inevitable. We can’t win.

After all, every image or video is just a set of pixels arranged on the basis of simple, binary signals. We’re already at a point beyond which humans are incapable of spotting differences most of the time, so any computer detection is unlikely to defeat a perfect machine-made forgery a few years down the road.

If anything we see on the internet can be made up and it is so easy and cheap to produce it, then it’s only inevitable that a wave of this fraudulent content is coming and we won’t be able to stop it just like Facebook can’t even stop bots from sending spam messages from single ladies in your area.

The obvious response to that is… a retreat from the internet (at least those spaces where it becomes a problem).

Back to basics

Ultimately, we may arrive at a point at which we will never be completely sure if there’s a real person on the other side unless we can sit in front of them, see them, touch them, hear them speak.

Simply put, use our basic senses.

It’s the only alternative to digital screens.

Seeing things for yourself, experiencing life, meeting friends, family, colleagues and strangers is going to be of much higher value than the internet where you can’t even be sure if the person on the other side is real.

Feed a machine enough information and it will be able to have a perfectly real conversation pretending to be your high school friend, your brother, sister even parent.

What if the “white lie” of the future is having a digital avatar to entertain your buddies while you’re away? Or to merely pretend you’re chatting with them when you don’t really give a damn?

Heck, Microsoft’s latest tool, VALL-E is able to recreate anybody’s voice from just a three-second recording — and then have them tell you anything (how long before scammers pick it up to do fake calls extorting ransom for a “kidnapped” relative?).

This is where it extends beyond just the internet. What if you couldn’t even rely on something as simple as a phone call?

AI can not only undermine huge parts of the internet, powering modern human interaction, but significantly cripple 150-year-old invention of the telephone, pushing us back to pre-digital era of communication that relies on direct interpersonal interaction.

Could you ever be absolutely sure that you’re speaking to the actual person and haven’t been rerouted to someone else trying to make you do something using attributes of someone you absolutely trust?

Singapore is already being plagued by scammers pretending to be policemen or Chinese officials, successfully duping people into giving up hundreds of millions of dollars each year.

And this is only from those technologically inept, easy to manipulate or simply gullible. But what if you picked up the phone and the voice on the other side sounded like your mother or spouse?

How many more can fall victim to impersonation crimes fed by machine learning models able to accurately mimic how someone speaks after recording just a short sample a few hours or days earlier?

People got all excited about working from home during the pandemic but the rise of AI may end those Zoom calls for good, given how even today people are trying their damnedest to pretend they’re participating when they’re not.

What if they could simply put up a completely realistic persona that looks, speaks and behaves like them?

It’s either the death of meetings (yay!) or, more likely, death of remote work (nay…). And it would be for your own good because if your employer figures out that your digital clone is about as productive as you are, then why would they keep paying you at all?

Education is about to take a step back too. Given how quickly students jumped on ChatGPT to do their homework for them, how will teachers be able to verify the skills of their pupils? Only in person.

Homework is not going away — just the “home” part. Instead of taking assignments out of school, it’s only natural to have them do more work in it, preferably away from computers.

Just a few years ago, tablets were touted as an obvious replacement for textbooks, but they might be gone before they even took hold of the market (or they might arrive with preloaded content and disabled internet connectivity).

Blessing in disguise?

But perhaps, we should be happy about it after all. Isn’t it the answer to all of the concerns raised about screen time and our over-reliance on technology, as our lives have become too digitised?

Forget Facebook, just go out with friends. Forget Zoom, meet with your colleagues instead. Forget Tinder, go back to the club and look for a date.

Technological progress in the domain of machine learning and AI may soon create a highly bipolar world in which we will heavily rely on this novel technology to improve our productivity but, at the same time, will have reverted to tech-free experiences outside of the narrow scope of work.

AI learning to do everything we do may just as well remind us of who we really are.

Featured Image Credit: Unreal Keanu Reeves via YouTube