How Bots Are Conquering the Internet

This illustration picture shows icons of AI apps on a smartphone screen in Oslo, Norway, on July 12, 2023. Olivier Morin/AFP via Getty Images

Commentary

Anyone who has performed a Google search, posed a query to ChatGPT, or made a purchase on Amazon has interacted with a bot—a piece of software that runs automated tasks. If you use digital technology, bots influence your life on a daily basis. Bot activity accounted for 49.6 percent of all internet traffic in 2023, according to a report by cybersecurity company Imperva.

Like bacteria in the human body, there are good bots that serve useful and even essential functions. Web crawlers such as Googlebot index websites for search engines, chatbots answer complex questions and draft emails, and transaction bots verify your credit card details when you buy a product online. Despite the service of these helpful bots, most bot activity on the internet is harmful. “Bad” bots accounted for 32 percent of all internet traffic in 2023, compared with 17.6 percent for “good” bots, Imperva found.

Bad bots engage in mischief—from extracting data from websites without permission to outright criminal activities, including fraud and theft. Scalper bots scoop up limited-availability items and resell them at a higher price. Cybercriminals use sophisticated, versatile bots for a wide range of nefarious activities such as credit card fraud and gaining illegal access to user accounts.

While typically not illegal, spam bots that push out unwanted advertising or backlinks are a major nuisance on social media platforms. Elon Musk threatened to walk away from his 2022 purchase of social media giant X, then known as Twitter, after alleging that the company was withholding data on the prevalence of bot users. The purchase was eventually finalized, but the bot infestation on X has persisted despite Musk’s pledge to “defeat the spam bots or die trying.”

Advertisers pay money to show ads to real human users, so the question of how many X users are actually bots is critical to the company’s business model. The true percentage is anyone’s guess. This author has had the experience of being swarmed by “fake” followers after setting up new accounts on X. These followers are easily identifiable as bots because they were all created in 2023 and 2024, they have unusual usernames containing strings of random numbers, they all have female profile pictures, and they never post any comments.

While the purpose of these unwanted bot followers is unclear, many social media users—including businesses, politicians, and entertainers—purchase fake or automated followers to boost their perceived social influence and online engagement. Bot accounts have also become a notable factor in political discourse, with political campaigns and foreign governments deploying armies of bots to manipulate online discussions and amplify certain narratives.

A study last year by Michael Rossetti and Tauhid Zaman analyzed the impact of social bots on X during the first impeachment process of then-President Donald Trump from December 2019 through January 2020. According to their analysis, bots posted more than 30 percent of content related to the impeachment despite accounting for just 1 percent of users.

Bots are also being mobilized to churn out low-quality content across the internet for a variety of commercial and personal reasons. A study by Originality.ai, an artificial intelligence (AI) and plagiarism detection company, found a 189 percent increase in AI usage on LinkedIn posts following the release of the popular chatbot, ChatGPT, at the end of 2022. As of October 2024, 54 percent of all long-form posts on the networking site are likely AI-generated. Likewise, the rise of AI tools that generate images from text commands has unleashed a flood of weird, deceptive imagery on social media platforms such as Facebook and Instagram.

The increasing prevalence of bot users and machine-generated content has fueled debate about how much of the internet is actually fake. The “dead internet theory,” which emerged in the late 2010s and has drawn renewed attention in recent years, posits that bots and AI have essentially taken over the internet, making it a synthetic, inhuman space where the vast majority of activity is driven by algorithms rather than people.

Proponents of the dead internet theory claim that most of the content, entertainment, news, and people that we encounter on the internet are fabricated by software—for example, that many YouTube personalities are in fact AI-generated videos, or “deepfakes.” While that probably isn’t true, the growing sophistication and rapid deployment of AI models make it easy to imagine a future scenario in which the internet is totally overwhelmed by worthless computer-generated content.

In a 2011 essay, Silicon Valley venture capitalist Mark Andreessen famously declared that “software is eating the world.” Increasingly, it seems that automated software programs are eating the internet. Bots, both friendly and malicious, drive half of internet traffic, giving them significant influence over trillions of dollars of economic value. Bots talk to us, compete for our attention, manipulate and propagandize us, and flood our websites with artificially generated text and imagery. As bots run rampant, it’s worth pondering the price we pay in human autonomy.