The Internet isn't dead yet

In the 1960s, Lake Erie was declared dead. The smallest of the Great Lakes (by volume) was plagued by algae blooms and low oxygen levels, killing fish and other critters. This environmental catastrophe had been caused by industrial pollution, largely from a lack of government oversight that allowed factories to dump their waste into Lake Erie and the rivers that feed into it. One of those rivers, the Cuyahoga River, became emblematic of this rampant pollution when it caught on fire in 1969.

In 2023, Emily M. Bender, Professor of Linguistics at Washing State University, described a different kind of pollutant in a different kind of ecosystem. She had shared an image of what was captioned as an adorable baby peacock. Wide eyed and with brilliant metallic blue and green feathers, she was soon informed that the image was generated by an AI image generator. When she did a Google image search for “baby peacock” among the legitimate photos of gangly little brown and white peachicks was a proliferation – an oil spill, as she put it – of the AI generated fantasy art that in no way resembled the reality.

The idea that our online information ecosystem has been polluted by junk information is not new. In 2011 it was content farms using human-generated low-quality content, later that worry was supplanted by bots proliferating nonsense across the internet, and today it’s the surfeit of AI generated content entering our searches and social media feeds. This flood of non-human generated, synthetic content has helped to bolster and popularize the Dead Internet Theory. However, declaring the internet dead may be both premature and counterproductive.

The Dead Internet Theory states that the internet is no longer a place where humans create content, but where any purported human-generated content is actually created by Artificial Intelligence. It plunges into conspiracy theory territory by inferring that the entirety of the internet is now a means to control and manipulate the humans that unwittingly consume content with all AI controlled by the government, corporations, and other iterations of “the man.”

The Dead Internet Theory keys-in on the very real problems of anticompetitive behaviour among tech giants, particularly those in AI, as well as the proliferation of synthetic content in our digital ecosystem. However, where the Dead Internet Theory is counterproductive is the nihilism with which it conceptualizes those problems. Proponents of the Dead Internet Theory see everything on the internet as being fake, with nothing being deemed trustworthy.

When dead fish littered the shores of Lake Erie, and photos of flames on the Cuyahoga River made the front pages of newspapers [1], people didn’t just ironically shrug and give up. Rather, a rush of environmental protection regulations during the 1970s and 1980s resulted in a decrease in pollution and increase in the monitoring of lake health. Today, the water quality in Lake Erie is far better than it was in the 1960s. The ecosystem is still fragile and continued vigilance is still required, but the littlest of the Great Lakes is far from dead.

Likewise, but with caution, regulation is needed to clean up the internet. Emily M. Bender, at a Congressional round table, stated “When Dow or 3M or Exxon or whomever spills chemicals into the physical ecosystem, we work to hold them accountable. We should build similar protections for our information ecosystem.” Morten Rand-Hendriksen, an instructor with LinkedIn Learning, is also a proponent of regulation but cautions that large companies can manipulate the regulatory process to their advantage if non-tech savvy policy makers fail to heed the advice of impartial experts.

Another issue with the Dead Internet Theory is that its key tenets are demonstrably not true. Yes, some parts of the web are littered with AI generated content. But the web still has human generated content and human interaction. For example, I am a verifiably real human-being with a web presence. You can read my articles here and on my blog at erinkernohan.ca, I post photos of flowers and bugs to Instagram, and I talk to other real humans on social media that I’ve met in person. You probably do too.

The assumption that there are no humans on the internet creates a dehumanizing effect. The more we believe who we are interacting with aren’t human, the more we stop treating each other humanely. Likewise, if we consider all content fake with no critical thought, then we lose a foundation of agreed upon facts and any measure of trustworthiness. We fall from healthy skepticism into dangerous cynicism.

We need to view AI-generated pollution on the internet the same way we view environmental pollution – as a problem we need to solve and can solve with humans and our common good in mind. And we need to act before something catches on fire.

Learn more

The Dead Internet Theory. Wikipedia. Last accessed 2024/04/20.

Maybe you missed it but the internet 'died' five years ago. 2021. Kaitlyn Tiffany. (The Atlantic) Last accessed 2024/04/20.

The Dead Internet Theory. 2024. Morten Rand-Hendriksen. (TikTok) Last accessed 2024/04/20.

Forget Crypto, Blockchain, NFTs, and web3: The Next Phase of the Web is defined by AI Generation. 2023. Morten Rand-Hendriksen. (Mor10.com) Last accessed 2024/04/20.

Cleaning up a baby peacock sullied by a non-information spill. 2023. Emily M. Bender. (Medium) Last accessed 2024/04/20.

Advocating for protections for the information ecosystem. 2024. Emily M. Bender. (Medium) Last accessed 2024/04/20.

Facebook asks for a moat of regulations it already meets. 2020. Josh Constine. (TechCrunch) Last accessed 2024/04/20.

Content Farm. Wikipedia. Last accessed 2024/04/20.

"Lake Erie is Dead." 2010. Michael Rotman. (Cleveland Historical) Last accessed 2024/04/20.

Great Lakes Pollution. (Michigan in the World/U Michigan) Last accessed 2024/04/20.

Correction log

[1] While the 1969 fire on the Cuyahoga River was indeed used as a call to action to regulate industrial pollution at the time, there are no known photos of this particular fire. A Time Magazine photo of the Cuyahoga River ablaze has been misattributed to the 1969 fire, and in reality is the much larger 1952 fire.

Yes, the 1969 fire was covered in Time Magazine, National Geographic, among other publications. But any photographs in existence are of the previous fires the river experienced.

Source:

Fables of the Cuyahoga: Reconstructing a History of Environmental Protection. 2002. Jonathon H. Adler. Fordham Environmental Law Journal Vol XIV. Last accessed 2024/04/20. [Archive link]