From its primordial days as a part of the United States’ military defense systems during the Cold War in the 1960s to the birth of the World Wide Web and Netscape browser in the early ‘90s, the internet’s primary purpose – at least for its first half-century – was largely pretty clear. It was made to keep people in touch with each other, a once-revolutionary global driver for human-to-human communication.
Many might still offer that raison d’etre if asked “what’s the internet for?” today. But something under the hood has been quietly changing – the internet’s humanity is quickly fading, and that fundamental change may just have crossed a vital precipice in recent years. The rise of the dead internet theory and its bot-driven internet landscape grapples with this future, and it might just change how we interact with each other online and online safety in general forever. Here, Spokeo explores what you need to know.
What Is Dead Internet Theory?
The dead internet theory first started popping up online in the late 2010s before really catching on in 2021. Its thesis is pretty simple: in the future (spoilers: potentially a very near future, or even the present), the internet will be more populated by autonomous bots than by human beings.
Who's Calling Me?
Search any phone number to learn more about the owner!
How this might happen – and how some might argue that it already has – isn’t explicitly agreed upon in the original theory. But we’re seeing it play out via a few broadly converging waves, some of which have been shaping the internet for decades, and some of which have transformed it rapidly in just the last couple of years:
- The rise of bots. Internet bots, automated software applications typically programmed to imitate human behavior, have been around since the early days of the internet, most notably used as social bots. These chatbots can engage in simple conversation with real humans, often deployed for customer service applications, entertainment, or to automatically propagate scams en masse, mostly via impersonation.
- In recent years, the proliferation of bots has converged with the rise of large language models (LLMs, or generative AI) such as ChatGPT, Claude, Gemini, Copilot, and Grok. These have all made bots more sophisticated, easy to produce, numerous, and autonomous than ever.
But LLMs aren’t just producing chatbots; they’re algorithmically churning out images, articles, videos, voices, music, and more. The mass creation of AI-generated content (colloquially called “slop”) is transforming the makeup of online content from human-made to AI-generated. - Simultaneously, human-made content produced pre-AI is slowly disappearing, as these older pages are no longer being maintained. This phenomenon is known as “link rot,” as older links are increasingly leading to dead content, leaving a void to be filled by autonomous bots and AI slop.
By the Numbers
Here’s the thing about dead internet theory: while it’s got “theory” right in the name, it does have the numbers on its side. Even before the term gained traction in 2021, data reflected a startling proliferation of online bots, and since then, the stats have only gotten less and less human.
- Cybersecurity firm Imperva’s annual “Bad Bot” reports clearly illustrate the trend, with bots making up 42.3% of internet traffic in 2021, 49.6% in 2023, and finally crossing the threshold in 2024, with 51% – the majority of internet traffic – generated by bots. By 2025, that number had crept up to 53%.
- Live global internet tracking service Cloudflare puts online bot traffic significantly higher in 2026, at 60.6%, compared to just 39.4% human traffic, as of mid-August.
- Data from the Pew Research Center estimates that 38% of human-made webpages from the year 2013 alone no longer exist. A quarter of pages made between 2013 and 2023 are no longer accessible.
- Likewise, 23% of news webpages and 54% of Wikipedia pages contain at least one broken link, while one-in-five tweets are no longer publicly available.
- On the social media front, an estimated 95 million accounts are fake, with 64% of all Twitter accounts estimated to be bots. About 18.2% of business account followers and 23% of influencer followers are likely low-quality, bots, or fake.
What It Means for Online Culture
Essentially, a dead internet becomes an online ouroboros – a digital snake eating its own tail— as bots both generate content and interact with that content. Early 2024’s “shrimp Jesus” phenomenon is a great example. As uncanny AI-generated images of crustacean-Christ chimeras spread throughout Facebook, people wondered how these bizarre pics routinely drummed up over 20,000 likes and comments. Turns out, bots made up the majority of that engagement – a memorable and highly visible case of bot AI accounts reacting to AI-made content.
As former Washington Post reporter Taylor Lorenz puts it, “I think the internet was terminally ill before ChatGPT was announced and released. Algorithmic ranking systems, which are AI-driven, really set the stage for just endless, worthless pieces of content and for the whole internet to be optimized in the most absurd ways.”
What that absurd, bots-serving-bots sort of content loop means for the remaining human denizens of the internet is seismic, but it can broadly be boiled down to two culture-shifting pillars: we can no longer trust that the content we see and read online is real, nor can we trust that the “people” we’re interacting with online are human beings.
Dead Internet Theory and Online Scams
A dead internet is more than just a lonely digital space made up of junk content and bot-driven engagement that looks good on paper for the corporations that stand to profit from it by juicing their almighty engagement numbers and ad views; it’s a space where actual humans become the target of explicitly malicious bots.
With the proliferation of bots inevitably comes the rise of “bad bots,” or bots with malicious intent such as phishing, identity theft, impersonation, and other scammy ways to get at the sensitive information, including bank accounts, of real human beings online. According to Imperva’s data, AI-fueled bad bots alone make up 37% of all online traffic, with bot account takeovers rising as much as 10% per year and more than 30% of API attacks being automated.
How to Spot a Bot
With up to 300 million fraudulent sign-in attempts per day, 43% of data breaches performed by bots, and over $5 billion in estimated annual losses from malicious bot attacks per year, identifying and blocking bad bots has become fundamental to internet security. Personal steps like enabling multi-factor authentication for logins and backend practices like advanced detection algorithms and verification tools (like CAPTCHA systems) help keep us safe. But often, it falls to the individual to identify bot content with common sense, a touch of media literacy, and our own eyeballs.
Here are some common red flags that scream “bot,” from social media chatbots to bad bots designed to impersonate professional organizations or otherwise catfish you out of your money:
- Generic, stolen (a reverse Google Image Search is your friend), or AI-generated profile pictures.
- Usernames that no human on earth would use, especially those with lots of numbers, like “molly0182789y.” Brand-new accounts with few posts and followers are also signs of bot accounts on socials.
- Bad grammar and odd phrasing (the classic “Best regards” email sign-off, for instance).
- Inconsistencies. Generative AI is based on algorithmic output; it tells you what it thinks you want to hear based on vacuuming up the bulk of content on the internet and spitting out the average of that content. That means it’s not great at keeping a story straight. If that slightly off “person” in your DMs or suspicious dating app match can’t seem to keep the details consistent when asked, it could be a bot.
- Strange or overly fast response times. Keep an eye out for the typing indicator – humans have to type messages, bots don’t.
- Instant social engagement, especially with repetitive behavior. See a social post with hundreds of immediate responses that are eerily similar (“Love this,” “Wow,” etc.) from a bunch of strange usernames? Probably bots. Likewise, high follower counts with low engagement can indicate bot followers.
- Behavior that pressures you to take action or creates a sense of urgency, such as urging you to click a link or make a payment. Be especially wary of threats (“if you don’t pay now”) or promises of unrealistic rewards.
- Off-brand behavior as well as incorrect URLs, social handles, and email addresses can be a sign of bots claiming to represent a business or organization.
How to Stay Safe
As ever, be skeptical of online behavior on both ends of the spectrum – both the kind that puts uncomfortable pressure on you or seems too good to be true (be that romantically or financially, especially). Block and report any bots you confirm, and don’t engage with bots you suspect, especially in the age of post-AI data scraping, where virtually any info you provide can potentially be leveraged for AI-driven identity theft. Enable multi-factor identification, vary your passwords or use a password aggregator, and use your devices’ built-in security features. Finally, never, ever click a link you’re uncertain of to make a payment to or provide sensitive or financial information.
Professor of Management and Digital Transformation at Kalaidos University of Applied Sciences, Dr. Yoshija Walter, writes that the dead internet theory “used to be rather speculative,” but “with the wake of generative AI, it can now be observed first-hand.” And while that inherently makes navigating the internet a more precarious and confusing prospect, knowing exactly what you’re observing is half the battle – at least, for us humans still left online.
As a freelance writer, small business owner, and consultant with more than a decade of experience, Dan has been fortunate enough to collaborate with leading brands including Microsoft, Fortune, Verizon, Discover, Office Depot, The Motley Fool, and more. He currently resides in Dallas, TX.