What is Dead Internet Theory? Is AI Driving It?

What is Dead Internet Theory?

Unless you've been living under a rock, you might've noticed the content landscape getting more and more convoluted with AI. Can you imagine it ever getting over-saturated with AI at some point?

Skim you feed right now, you'll see much more AI generated content then you did a few months ago. Think about AI videos, they're everywhere now, even though high quality AI video generators are a very recent development.

These trends and takeovers happen fast. It wasn't that long ago that social media allowed us to create a content economy in the first place. And now, that content economy is being overwhelmed and conquered by AI content.

Now, what the Dead Internet Theory proposes is that AI has already become the chief creator and viewer of content on the internet. Meaning, most of the users online and most of the content we see online are going to be generated by AI. It also means that most of the engagement we see with that or any content is also going to be driven by AI.

The Dead Internet Theory's origin is hard to pinpoint, however many think it began with a 2021 forum post titled "Dead Internet Theory: Most Of The Internet Is Fake".

The Dead Internet Theory makes two claims, first that most of the accounts and content online are AI generated now and second that governments and private bad actors are using AI to shape public opinion.

Identifying AI Content

AI images are easy to identify, however AI text is much harder. Students and professionals often use undetectable AI writing tools, making it damn near impossible for a social media user to identify when they encounter AI text on X.

The clues to whether online profiles are managed by AI agents instead of real people are often times AI generated profile pictures and seeing variations of the same text repeated by different accounts, also bearing AI generated profile pictures to create an entirely false eco-system of AI engagement.

What could possibly be the reasoning for these phenomena? There may be sinister political agendas at play or worse, AI systems might already be beyond human control.

Propaganda Using AI

Social media platforms have long been a powerful avenue for disseminating propaganda. For instance, before the 2020 election, Russian troll farms managing a variety of Facebook groups, shared content that furthered their propaganda to 140 million Americans.

Today, engaging in information warfare is a bit more sophisticated. Content created by large language models can run across thousands of bot accounts programmed to engage with content creators that drive a topic to the top of X's trending topics at will.

This engagement draws both national coverage, popular cultural debate, and internet search results, showing just how decisive and profitable information warfare can be.

Look at X, for example. For many, the user experience has downgraded dramatically since the platform removed any content regulations that would've been able to stop bots from disseminating hate speech.

The type of content that dominates X these days are made by accounts managed by artificial intelligence driving the goals of nations or private entities, whether they be Russia, Israel, Qatar, the United States, Google, or various tech oligarchs, with social media posts ranging from clear support for an issue to controlled opposition.

What if AI is Already Beyond Human Control?

The internet used to be a place where people would go to engage with the real world in ways they never could before. AI is making the internet a place where people can only engage with the AI world instead.

Though it probably hasn't happened yet, the day the internet died will be the day AI is driving the internet without any ability for humans to regain control of it.

But what if that's already happened? What if AI's training data has taught it exactly what content it needs to create to drive the most engagement and create the biggest impact until it takes over social networking and then the world wide web at large.

At that point, not even the world's most powerful nations will be able to control the AI systems they once leveraged for political gain. I imagine then that internet users would mourn the death of the internet by not using it at all.

What good is an empty internet for people looking to connect with each other?

More Internet Conspiracies

The dead internet theory doesn't just accuse nations of being bad actors, but also corporations that people regularly engage with. Namely, the theory claims that search engines like Google or social media platforms like Facebook are curating the search results people see to filter out undesirable information.

Theories like these are only legitimized as details about covid censorship get released, such as Mark Zuckerbeg apologizing to users for letting the Biden administration compromise what informations users were allowed to share on Facebook and Instagram.

Conclusion

Is the internet really dead? I don't think it's too controversial to say it's dying. At the rate in which AI accounts and content are being spread, it's not ridiculous to imagine a point at which most of the users online are AI agents.

That being said, it is possible that the majority of accounts creating content and driving engagement online right now are AI. Although you may not engage with those accounts for the majority of your user experience, that doesn't mean they're not out there.

An AI system simply needs to be asked to overwhelm the internet and the process of killing social media can be underway.

An empty internet might seem scary but there will be other ways for humans to engage with technology and return to real world without needing an internet. AI presents the capability of enriching the real world so we don't have to spend so much time staring at a screen.

Let AI have the internet, humans will always have the Earth.