The Undead Internet: The Web of the Living Dead

Jonas Hultenius

2024-11-01

The internet, once a vibrant and dynamic space for human interaction and creativity, is undergoing a transformation. And since it’s Halloween season it’s all too fitting that it is a scary one as well.

Over the last couple of years, it has been evolving into a realm increasingly dominated by artificial intelligence, where machine-generated content is blurring the lines between the authentic and the synthetic. I talked before about the dead internet theory and how the web now more a hollow shell of what it once was and how algorithmic content is taken hold. Well, we are well passed that point by now, the internet is no longer dying, it is undead.

For years algorithmic content has been summing and regurgitating content, leaving the web a digital landscape devoid of human creativity and originality. A shell of its former self. But something has changed. Now it’s filled with new content, original, beautiful and completely fake. A Frankenstein monster, new original content made in its creator’s image, out of the remnants of the dead web.

I told you it would be scary!

At the heart of this transformation is the rapid advancement of AI technologies, particularly large language models. These models are highly capable of generating human-quality text, writing code, translating languages and creating new and semi original artworks. So good that we often are fooled now a days and not being able to tell what is fake from what is real.

While this technological progress offers immense potential, it has the potential to unlock endless creativity and supercharge human creations, it also raises concerns about the impact on the quality and the authenticity of online content.

As AI-generated content floods the internet, it risks diluting the quality and originality of the content we consume. Search engines, social media platforms and content aggregators are becoming increasingly saturated with AI-generated articles, blog posts and social media updates.

This leads to information overload in general, but also making it increasingly difficult for users to distinguish between human-generated, real and AI-generated content. Quite often that is not a bad thing. Most things can be interchanged between real, surreal, completely synthetic and partially real. If a product page displays an ice-cold beverage in a red aluminum can with an iconic logo it does not really matter if the can is real or not. If it is a genuine photo, something created by hand or simply generated does not impact what it’s portraying.

For other content however, authenticity is much more important. Just a few years back, in the age before the current AI boom, websites and marketing material had to be handcrafted and there were clear signs that something was not quite right. Companies could create filler by using stock photos and cobble tougher an offering by cutting and pasting from others. This often gave a good impression at first glance but there were warning signs and dead giveaways.

Now anyone can create a great-looking website with near perfect copywriting, without misspellings and with a certain panache and flair that we associate with well-managed and reputable companies. Product pages, which once had to be filled with genuine articles or stolen images, now can be generated for products and services that do not and might never exist. All with a fidelity that is above and beyond anything we could imagine just two years back.

I could set up a website selling jeans in a matter of a few hours with a consistent messaging around them, great looking products and models that fits my brand wearing them and customer endorsements talk about our quick deliveries and customer service center. The only thing remaining is the sourcing of the actual garments and even that is optional.

The web is overflowing with beautiful images that are near real but depict things that never existed. And while that is phenomena is exciting to me things are quickly getting out of hand. If you for example do an image search for ‘Baby peacock’ you quickly find amazing images of tiny, cute and colorful birds. Birds that never have existed and that do not even resemble what the actual real birds look like in their juvenile appearance. It turns out that they are really bland and boring at first. Who could have known?

These are all signs pointing to the internet dying but I would go so far to say that we are entering into the age just after death, and the reanimation of the web into something else.

The initial AIs, our synthetic helpers and friends, where done on the enormous backlog of information that we have put tougher since the early nineties the new generation of AI is trained on the current state. The AI is being trained on synthetic data to become even more synthetic and even worse, we are being trained on that same data as well.

Untill now the internet has been a place of truth or lies, and to some extent half-truths and misconceptions. There has been some level of checks and balances between the two. Great looking websites cost money and cheaper looking once are, or were, often scams. Content cost money and having a lot of it, and high quality, would indicate a sizable investment. An investment that is out of bounds for most conspiracy theorists, scammers and tricksters without connections to a shadowy cabal or foreign nations secret service of course.

The potential for AI to generate vast amounts of misinformation and disinformation is hugely pressing concern. With AI-powered tools capable of creating highly convincing fake news articles and social media posts, it becomes increasingly difficult to discern truth from fiction. This can have serious consequences for society, as it erodes trust in institutions, fuels political polarization, and undermines the democratic process. And this has been for debate for some time, making the term ‘fake news’ a household name used by even my own children.

And while it is a true problem that needs to be addressed, the even more pressing subject, to me, is that the internet itself might be in danger. When we can’t even truest that images for baby peacocks are real, what is?

We all have a notion of what is real and not and that commercials with astronauts drinking cans of soda in space might not be one hundred percent real. And that news and certain facts might be twisted to better fit into the narrative. But when everything digital might be fake, and good fakes at that, what should we believe.

While the potential benefits of AI are undeniable, it’s crucial to recognize the risks and take steps to mitigate them. Transparent disclosure of AI-generated content is essential to maintain trust and accountability. AI tools should be used as aids to human creativity, not as replacements for it, and when used we should be upfront about it. There is no shame in it, they are a powerful tool to help us get our ideas across and generate new once.

The internet today is more and more a reanimated gestalt of what it once where. A zombie brought back from what remains from what came before. Ultimately, the future of the internet depends on how we navigate these new challenges. AI is a tool for enhancing human creativity and problem-solving, not a replacement for it.

There is still time to slay this beast and ensure that the internet remains a vibrant and dynamic space for human interaction, innovation and creativity. We just need to figure out what authenticity is first. Happy Halloween!