Could the Internet Become Useless?
Much of the internet is a mess. It's going to get worse. Then it will likely get better. But why and how?
While estimates vary, it is theorized that somewhere around ~45% of internet traffic is largely just junk. Spam, scams and just meaningless bot traffic. Now, researchers are seeing the rise of Generative AI content, known as slop. Some have suggested the Internet might become largely useless. Could it?
Yes. And no. But it could become very different from what it is today. But that always happens with technologies. When penny dreadfuls came out in the mid 19th century, along with other cheap novellas, many a pundit suggested it was the end of literature. It was not.
The internet will likely become much better, more useful, more valuable across all aspects of society. It’s the getting there that’s a bit wonky.
When the telegraph burst it’s way into the world, some believed we would end up talking in staccato sentences and poetry was doomed. We have a long history of overestimating the impact of a technology in the early days, known as Amara’s Law. But the internet, at a global scale, is a few decades old. Why the angst of a useless internet?
It’s worth considering as the internet has become deeply enmeshed in the very fabric of our global sociocultural systems. The internet impacts all aspects of culture; the aesthetics (music, art, literature, fashion), economic and political systems, militaries and social governance. How it evolves now impacts humanity at a species level.
This in part, is also why coders should not be the only ones who decide on it’s evolution. The internet is a general purpose technology. To be of benefit and value to humanity, it means a whole of society approach to its future.
Could The Internet Become Useless?
While that is highly unlikely, we may well go through, and perhaps are already in, a phase where things get very messy, complex and confusing. This is not unusual and we’ve always figured these challenges out. As electricity became a reality, it upended economic models and changed even how we design cities. Factories no longer needed to be situated near rivers and streams. This enabled the creation of industrial parks and outlying suburbs alongside the development of the automobile, rail systems, airports and so on. There were many societal fears of electricity.
The challenge with the internet is one of both infrastructure and the information that flows through that infrastructure. Lets look at both. And why Artificial Intelligence (AI), our newest, potentially general purpose, technology needs to be considered.
Infrastructure: The internet (being synonymous today with World Wide Web) to most people, is both very robust and very fragile. A backhoe doing some boring digging for a road can snap a fibre optic line. A hostile government can snip an internet pipe with a submarine or a dive team. Servers can become flooded and overwhelmed with a DNS attack. Cybercrime and cyberwarfare.
Artificial Intelligence: Beyond the issue of poor and often meaningless Generative AI (LLMs etc.) flood of slop content and spam, is the massive energy cost of the data centres needed to make it all work. These datacentres use vast amounts of water and electricity. The planet does not have enough of both to meet the projected demands of the AI industry. We must also consider the datacentre demands outside even just AI.
The Information: This is the world of mis/disinformation, streaming content, all the massive volumes of human generated content beyond cybercrime and spam. The good stuff we create. Much of it today, is stored in the Cloud, a metaphor for large datacentres.
It was largely believed in the mid to late 1990’s that the arrival of Information Communications Technology (ICT) would result in a huge uptick in productivity. Instead, we’ve been locked in what is called the productivity paradox for nearly thirty years with, no idea how to get out of it.
The assumption has been that GAI and other AI tools would solve for this problem. So far, it is making it worse, not better. As GAI tends to hallucinate, it means having to double check everything an LLM creates. For many businesses this means increased legal fees and more work for marketing, not less.
Google has become the standard bearer for organising the world’s information. It is now struggling to even maintain the quality of basic search functionality. A problem it is throwing vast resources at. In the online advertising world, MFAs (Made-For-Advertising) sites, clickbait and scams are showing that between 15–50% of advertising budgets are wasted.
What Are We Going To Do?
There’s no simple or easy answer. This is a complex, multifaceted set of problems. Unfortunately, AI tools so far, are making the situation worse, not better. In time, they will hopefully improve enough. For now, they’re just adding noise and confusion in an attempt to stick a band-aid on a sucking chest wound.
We are likely to see it get worse in the short term. There is no singular solution, but rather lots of small and big technology fixes over time. But just using technology to try and fix all these challenges won’t work either.
Social media platforms believed they could solve the issue of spam and dealing with mis/disinformation with AI tools. That failed miserably and they’re still using thousands of humans to deal with the issues. This is the evidence we should use to understand a whole of society approach is needed to be part of the solution.
This means applying the laws, regulations and processes such as ISO and IEEE standards to the digital world. Something we’ve not done a very good job of so far, but are beginning to. We see this through the EU’s Data protection acts and GDPR for privacy, California’s strict privacy laws and the Province of Quebec in Canada getting way ahead of the Canadian federal government. Lawsuits by State Attorney Generals and school boards against social media platforms.
Moving fast and breaking things, well, broke things. Societally. A rabid bull in a china shop on steroids. To prevent the Internet from becoming useless, we must think globally and socioculturally, not just economically or technologically.
The good news is that we’ve done this before. With automobiles, manufacturing quality standards, aeroplanes, telecommunications systems and so much more. While some in Silicon Valley believe that regulations and standards hurt innovation, reality has shown that belief to be false.
AI may well help us, especially areas like Machine Learning, Neural Networks and sometimes, Generative AI tools, but will not solve the myriad of challenges ahead of us. No, the internet will not become useless, in fact, it will likely end up becoming far more useful and valuable. History doesn’t repeat, but it often rhymes.
An excellent overview of an on-going problem. Thanks!