Over 30% of Web Traffic Comes From Bad Bots

Why Trust Techopedia
KEY TAKEAWAYS

  • AI bots are on the rise, presenting both benefits and challenges for online security.
  • Malicious bots can disrupt services, steal data, and manipulate online environments.
  • The fight against bad bots requires a multi-pronged approach and collaboration on regulations and ethical frameworks for AI development.
  • We speak to experts about how the future of the internet hinges on our ability to harness the power of AI responsibly.

Artificial intelligence is driving a massive surge in sophisticated online bot traffic. With studies warning that there is one bot for every human online and automated AI-driven bots generating a large percentage of global traffic, concerns begin to rise.

While not all bots in the digital space are “bad” — some are even necessary, such as

search engine crawlers that index websites or chatbots that provide customer service — malicious bots are growing to outnumber these.

But what are these bad bots up to? Why is there such an overabundance of them, and how does this impact the internet? Techopedia talked to experts to answer these and other questions.

CrowdStrike Reports Warns of Bot Dangers and Attacks

The June 24 CrowdStrike 2024 State of Application Security report found that a significant portion of the web traffic comes from potentially malicious bots.

CrowdStrike explained that malicious bots are often used to exceed the resources of dedicated application security teams. Bad bots can steal data, spread spam, launch denial-of-service attacks (DDoS), and much more. CrowdStrike warns about the danger of underestimating bad bots:

“Bad bots—if left unchecked—can cause massive disruption: One-third (31.2%) of all traffic stems from bots, the majority (93%) of which are unverified and potentially malicious. Top targeted industries were Manufacturing and Consumer Goods, Cryptocurrency, Security and Investigations, and U.S. Federal Government.”

Blocking AI Bots Has Become the Norm

On June 25, the popular social news and forum website, Reedit, announced it was taking further actions in its fight against AI bots. As generative AI continues to eat away internet data, Reddit announced it will continue rate-limiting and/or blocking unknown bots and crawlers from accessing the platform.

Reddit is not the only company that has banned AI bots, with most top U.S. news sites and big companies banning popular chatbots like ChatGPT due to security and privacy concerns.

But AI chatbots are only the tip of the iceberg when it comes to automated malicious bot activity.

There Is One Bot for Every Human On the Web

The 2024 Bad Bot Report of Imperva found that almost 50% of internet traffic comes from non-human sources. This represents one bot for every human online. Out of all traffic online, the report continues, 30% is malicious bot behavior.

Bad bots have become more advanced and evasive and now mimic human behavior in such a way that it makes them difficult to detect and prevent. They can also be leveraged in zero-day vulnerabilities, DDoS attacks, application programming interface (API) and firewall attacks, automated spam and phishing, and much more.

The Dystopian Dead Internet Theory Comes to Life?

The rise of AI bots, automated misinformation on social media, AI-generated content, and fake bot traffic have brought reminiscences of a bizarre conspiracy theory: The Dead Internet Theory.

Erik Severinghaus, Founder and CEO of Bloomfilter, an AI-driven software development company, summarized for Techoipedia the Dead Internet Theory in one sentence

“Ever feel like your social media feed is too polished to be real? Spoiler: It might be because bots are running the show behind the scenes.”

Believers of this wild theory claim that the Internet is dominated by bot activity and automatically generated content that manipulates search engines and social media algorithms for top positions. According to the theory, these bots were intentionally created to manipulate the population.

Severinghaus explained that the internet — which used to be very busy with people talking and sharing things — has now changed into a more complicated place where machines do important work.

“The Dead Internet Theory is very interesting but also a bit spooky. It says many things we see online are made by bots, not real people.”

Severinghaus recognized that the theory has some good points but is not entirely on target.

“Automation really changed how we use the internet, but it is just normal growth in technology,” Severinghaus said. “Instead of thinking this is the end of human-made content, we should see it as a change in how content is made and enjoyed.”

Bot Wreaking Havoc in Marketing, Malversting, Ad Fraud, and SEO Tactics

James Taylor, SEO Strategist at Embryo, an independent search-first marketing agency based in Manchester, U.K, spoke to Techopedia about the current bot situation in his line of work.

Taylor said that in digital marketing bots that generate vast amounts of fake clicks on Pay-Per-Click (PPC) ad campaigns have become a real concern. “At a time when budgets are already stretched, now marketers have to deal with ‘bad bots’ skewing data while draining advertising budgets.

“It’s costing businesses money while at the same time making it harder for them to measure the effectiveness of their SEO campaigns and taking up a huge amount of time in trying to counteract and remove the bots.

“The internet isn’t ‘dead’, but it is changing, and we should all start being a bit more critical about the content we consume and share. As the internet evolves, we are also going to have to evolve in how we use it.”

Taylor believes that while it may feel like AI is taking over, there is still a lot of human-generated content online. Additionally, Taylor explained that when humans compete with AI bots they might be pushed to improve the quality of human-generated content.

“That said, bots are definitely presenting challenges and whereas creating and manipulating bots once at least needed a certain amount of knowledge, AI is making it easier and easier for malicious actors to take advantage.”

Where Bots Do Real Damage

CrowdStrike’s latest report warns that the online world runs on web applications and APIs. The company that claims to block an average of 209 billion cyber threats every single day says that developers are not creating web apps with security in mind.

CrowsStrike found that bot-driven DDoS large-scale attacks have become the most popular vector of attack for cybercriminals. In these attacks, bots target web applications, firewalls, and APIs going after industries like Gaming and Gambling, IT and Internet, and Cryptocurrency, among others.

Severinghaus from Bloomfilter said bad bots are the “digital troublemakers”, revealing a dark side of automation.

“They [bad bots] copy content, take data without permission, start DDoS attacks to disrupt services, and spread false information. These bots cause many of the spam, fraud, and cyber threats we see on the internet.”

For example, bad bots make fake social media profiles to share false information and change how people think. They also carry out cyberattacks that can seriously harm websites and online services, leading to big money losses and damage to reputation.

“Fighting these bad bots is a nonstop problem needing to always be alert and using high-level security methods,” Severinghaus said.

“These advanced bots can act like real people very well, avoid security systems, and carry out complicated attacks that are difficult to spot. AI-powered bots have the ability to create believable fake news, influence financial markets, and conduct sophisticated phishing attempts.”

The Bottom Line

The rise of AI bots presents a complex challenge for the future of the internet. While “good bots” can streamline processes and enhance user experience, “bad bots” threaten online security, authenticity, and even democracy. As AI advances, detecting and preventing malicious bot activity will require continuous innovation in security measures.

Addressing the bot problem may necessitate collaboration between governments, tech companies, and security experts. Clear regulations and ethical frameworks for AI development are crucial.

The “Dead Internet Theory” may be an exaggeration, but it serves as a wake-up call. The internet is evolving, and we must adapt by being active participants, not passive consumers. The future of the internet hinges on our ability to harness the power of AI for good while mitigating the risks posed by bad bots.

Related Terms

Related Article