The Dead Internet Theory may not be a Theory anymore

What happens when theories aren't theories anymore?

Logically, we look for practical implementations to validate our new understanding, and if sanity prevails, we draw our conclusions from the evidence. Sounds straightforward enough, right?

The larger problem here is the validity of the understanding and how it has evolved over time, reducing it to a mere debate of my opinion versus yours. Facts don't seem to matter anymore.

Does AI deal in Facts?

The way AI models are trained, and I prefer not to use the term 'AI' in favor of 'LLM' or Large Language Models, is purely based on the quality of the underlying data. The data that is fed to the models under training can be entirely fictitious or based on real-world statistics, which cannot be confirmed by the end user unless they know what they are looking for when they use the models.

You see the problem here?

The LLMs can confidently serve wrong answers to the users because, at their core, they are trained on this very data. If the user is unaware of the basic facts on the topics they are researching, it is very easy for anyone to take the word of AI as gospel and base their thesis on it, and create a much larger false narrative that sounds like cold, hard facts.

The way LLMs are trained, which is by scraping existing data, results in false narratives being out there. The more it gets ingrained in our lives, the newer models are trained on it, where eventually it might even sound like the ultimate truth.

Sounds scary, right?

The Dead Internet Theory is based on Bots vs Human Interaction

If the above scenario sounds scary, imagine the recent push to use AI in every aspect of our lives and churn out content that sounds eerily similar without any personality to it.

The famous platforms like Twitter/X, Facebook, Instagram, Threads, TikTok, and many more promote the creation of content that is likable to their algorithm rather than the users. Hence the creators get sucked into creating content that is based on certain template, easy to be picked up by algorithm and hence easy to recreate. This results in brain-dead content that can be easily created by automated bots by repurposing existing content, which will be obviously more suitable for the algorithms on the platform to evaluate and likely to go viral.

It clearly replaces the end user from Humans to Algorithms, and thus the whole internet becomes a cesspool of content that is created by the machines to please the machines. Thus giving rise to the dead internet where humans no longer matter and get marginalized in the world of connectivity.

You can look up the definition and more material here -

The link has some great content on the theory as well as resources on how to use GenAI.

In case you prefer a quick video essay -

The Cause for Virality outweighs the Cause for Concern

Content Creators are just like any of us, trying to make a living by doing a job where the rules change dynamically and the goals are unclear or obtuse.

Automating the content creation allows them to try out different kinds of content in a rapid manner to test and push for the kind that works with the algorithms. They are not only competing against each other but trying to appease the algorithm as well, which is entirely proprietary and under the supervision of the creators of the platforms.

Soon, the need to use readily available tools without checking their relevance or integrity to create content that is not fact-checked becomes imperative in the race to be the next best influencer. For example, here is the post from 404media that explains Facebook's role in such kind of content creation -

The good old days of the internet being one large forum to research and a web of Instant Messengers are long gone. Instead, we have the SEO driven, AI-enabled content that thrives on human attention and monetization without any inclination to actually serve the basic purpose.

Where do we go from here?

All is not doom and gloom as the title suggests, but there are slivers of hope in the corners of this old internet. Before we abandon everything for the new internet as promised by the Blockchain community, we have a few tricks worth a try up our sleeves.

Humans, as we know, have this brilliant ability to create stories that are unique and reflective of their personalities. While the world sprints towards the adoption of AI in every facet of life, we would do well not to let it drown out the personalities that tell unique stories.

Algorithms tend to look at everything as black and white, and in turn, resolve everything to a statistic or a number. Life, as we very well know, is many more shades than just black and white.

The real stories worth reading and the work that changes the world for the better are always done in a spectrum of facts and emotion. Before we all start blindly trusting the LLMs with answers to our every question about life, source materials are waiting to be explored in the form of whitepapers and books that are vetted by people who have spent their lives on the subjects.

Use the LLMs to give you the source rather than relying on them for the answers. Use the source to formulate your hypothesis, the internet might be able to revive after all.