AI Paranoia

I have a visual impairment, it's caused by a condition known as Nystagmus. The short explanation is that my eyes constantly move involuntarily and because of this it is difficult for me to focus on fine details, the further away they are the harder it becomes. With most things close up I can manage quite well, this is something I have lived my whole life with so as you can imagine it's something I've learned to compensate for or find workarounds.

Computers, and technology in general make my life easier, the fact you can change the resolution of a screen and scale fonts to a comfortable size makes a world of difference for me. Existing in a world without the internet when I was a kid was very difficult, accessibility was always an afterthought, arguably it still is, but the assistive technology that exists today to aid me is designed to work with products and services that didn't add accessibility, but that technology enables it nonetheless - magnifiers, screen readers, audio transcription, larger screens, lower resolutions, adjustable DPI to name a few.

Using the internet when I was younger wasn't so easy before browsers added these abilities as features or before operating systems incorporated accessibility features on a native level that actually functioned. In the days of yore, I relied upon expensive third-party software which thankfully I didn't have to pay for as there were government grants you could access at the time here in the UK to gain access to the technology.

There was a brief "golden age" of accessibility for me sometime between 2006 and around 2016 but gradually spam became more of an issue for websites so they had to develop ways to try and prevent it. Sometime around 2010 the use of Captchas reached a point where I started to have personal difficulty trying to solve them. That only grew worse in time, I hated them with a passion, there was even a post I made on an old blog where I ranted at length on how irritating they had become.

Then the technology moved on, spam bots became more sophisticated and over time using Optical Character Recognition [OCR] the bots themselves were able to solve these Captchas - meanwhile I, too often, still could not. That was the first point for me where I realised that technology designed to prevent non-human activity online was discriminating against anyone who didn't fall within a narrow definition of what is expected of a human.

Today, the types of verification that exist on website vary, from sliding jigsaw puzzles that make you place a piece, to 3D models that ask you to rotate the image until it matches a request, the tasks people are being asked to perform to prove they are human are becoming more and more complex in terms of computation but again ostensibly these are tasks that "normal" people should find easy. I hate the word "normal" because I don't think it exists, it's ultimately an arbitrary statistical division and every definition of it relies on some degree of prejudice and discrimination to define it.

Spam is still a problem but it has evolved, Slop has taken its place - content which is either nonsensical or illusionary, generated by AI. Nonsensical in that it doesn't make any sense at all, or illusionary in that it seems logical at first but examine the claims it makes and none of them hold up - like the New York Times Summer Reading List article that contained books with blurbs that didn't actually exist.

The problem is veracity, defined as conformity to facts, accuracy - AI in its current form does not possess the ability to verify and validate what it claims, it relies upon its datasets and information scraped from the web, but anyone can contribute to the web, including AI itself, which creates an endless source of data to process but with no reliable means to verify any of it. Even as humans this is something we have struggled with, to the extent that certain websites gain a high value of trust - Reddit and Wikipedia as examples - and become the go-to destinations for answers with authority. The trouble with that is once again anyone can contribute to either site and if you've ever actually contributed to either you would know the community that exists on both controls much of what you are allowed to contribute.

Over zealous editors on Wikipedia protect articles from amendment, often themselves using bots to monitor the articles for changes and reverting anything they don't agree with. The idea that Wiki is not biased is something only held by people that have never actually edited it and ventured into the ecosystem within it and observed the inner machinations. The same goes for Reddit, each subreddit is a community in itself and if you invest any amount of time any those communities the same patterns emerge, from subreddits that require every new post to be moderated, to subreddits that allow posts by anyone but again use bots to auto-moderate and remove what they don't agree with, the same biases ultimately become entrenched.

Veracity has been a problem for as long as the general public has been able to contribute to the web, the only thing AI has changed is that the veracity of the data was the core focus before, but now the veracity of the contributor is as much, if not a greater concern. Putting aside the Dead Internet conspiracy theory for the moment, and focusing on the practical elements from a point of pragmatism, the ability to know what you are reading was written by a human is something we can't adequately implement. Social Media is the most egregious example and unfortunately part of the reason we can't adequately implement a solution is because verifying a human identity conclusively, ultimately requires identifying the individual, and most users, myself included, aren't willing to give that much detail about who they are, to prove they are human. I don't want to give a social network my full legal name, date of birth, address, or any other information that is so specific to me.

AI Paranoia is becoming a problem, circling back to the Dead Internet Theory, whilst the original post that spawned the theory is eccentric at best, the over arching idea, that algorithmic interaction now outweighs human interaction by several orders of magnitude seems not just plausible but seems patently obvious. Google's engineers working on Youtube refer to this as "inversion" the point where more views of a channel are from bots than from humans, when you look at the number of subscribers channels have versus views, that seems to be patently true already.

As an example, Mr Beast has according to Youtube, 400 million subscribers. I don't believe that figure, I am more inclined to believe around 20m to 30m people are subscribed and the rest are bots. My basis for this is an old metric I have often cited called the 90-9-1 rule, a metric that existed before AI and spam became prolific. The rule states that any community online generally has 90% lurkers, people who never interact, 9% occasional contributors, who only interact occasionally, and 1% active contributors, who contribute frequently and account for the majority of the content within a community.

On a recent video by MrBeast 'I Explored 2000 Year Old Ancient Temples' there are currently approximately 75,000 comments. Assuming that figure represents interaction of 10% (9% from occasional contributors and 1% from frequent contributors) with 90% of the people who viewed the video not commenting at all, then the real view count for that video would be around 750,000 views. If you take the same methodology and take views of that video as representing 10% of his subscriber base, again assuming 9% are occasional viewers, and 1% frequent viewers, with 90% absent not watching the video at all, then his subscriber count would be around 7.5 million. Roughly 1/53 of what Youtube says his subscriber count is.

If we reverse the process and take his subscriber count at face value, and it really is 400 million people, then 10% should be 40m+ views per video which to be fair looking at his channel his videos do routinely exceed that, but again 10% interaction should see comments around 4 million per video - none of his videos have interaction anywhere near that level. The video above with 75,000 comments has 85 million views at present meaning 0.09% of his viewers actually interact. That's 9 people for every 10,000 viewers. Imagine a football stadium filled with 10,000 people, all silent, in that sea of people there is one small group of just 9 people sitting together cheering. It's not hard to see why people think the internet is "empty" and void of real people.

Now to be clear I don't think real people have stopped using the internet, I think the same amount of people are here as has been for some time. The problem is that everything online is measured in metrics and from a business perspective the higher the metrics the more value you get to claim a business has. Youtube relies on ad revenue to be profitable and most of that ad revenue is derived from views on videos. Youtube tweaks the partnership and AdSense agreements ad infinitum to try and limit payouts to content creators and maximise the revenue received from paying advertisers, there is therefore a business case for complacency when it comes to inflated view counts.

Youtube isn't the only culprit, there was a time when you would google any search term and atop the results there would be a strap line which read "showing results 1 to 10 of 2 billion" or something to that effect - Google quietly removed this, along with the ability to "deep dive" into results - it used to be possible to go hundreds of pages deep into the search results Google served you, if you Google the word "Veracity" as an example Google will now only let you go 18 pages deep before it cuts you off, serving you approximately 180 results. Google has become surface and base.

Google search isn't just complicit in allowing this problem to grow, it may even become its undoing. Several years ago Google Inc became Alphabet Inc, something which some questioned at the time but I and many others pointed out that the real reason they did this was to restructure the business such that individual products could fail and be allowed to do so without jeopardising the company as a whole. AI wasn't seen as a motivating factor at the time because the general public weren't really conscious of AI and its implications, but in hindsight it looks increasingly like Google knew this was coming and rather than try to stop it they chose self-preservation through self-sacrifice - or to be more precise, through amputation.

The creation of Alphabet Inc facilitated the possibility that Google Search as a product could one day fail and Alphabet as a company could continue. This was something few people thought would ever be a plausible reality because rather infamously Google was "too big to fail" as many claimed. Today however Google itself has quietly admitted the utility of its search results are jeopardised by the veracity problem. The incorporation of AI summaries at the top of search results wasn't an attempt to overcome this problem but rather an "if you can't beat em, join em" moment, recognising a growing number of particularly younger users have been using ChatGPT as a search engine - something which many people are rightly horrified by but arguably that feeling of horror is belying a deeper problem - that Google never verified search results to begin with.

When Google was created, core to its functionality was an algorithm and metric known as PageRank. The idea was simple and novel at the time, to score each web page based on the number of pages that linked to it, and to treat most-linked pages as more credible than least linked, and to extend those metrics to transitive links (A to B to C) and order the search results based on those metrics. Over time the more popular a site became, the more others linked to it, the higher it rose in the search results. At no point did Google attempt to verify the actual content of these pages, truth was never something Google cared about.

PageRank evolved over the years, it became increasingly complex and Google added many features incorporating its own products and acquisitions, and it eventually caught the attention of regulators, some might say the ire, but even then the efforts Google made to "fact check" its results were cursory, often extending only to the headline result or definition or summary that Google itself offered, leaving the remainder of results open to manipulation. Even now Search Engine Optimisation [SEO] the process of optimising a web page for indexing by search engines still offers a wide array of tools that make it relatively easy to get any arbitrary web page to rank top of Google's search results, maintaining that position over time was difficult as you were competing against other humans optimising their pages, but now AI tools automate this process and they make it even easier to achieve those rankings.

AI hasn't eroded the veracity of results listed on Google, it has increased the volume of content that Google has to index and try to determine whether it was written by a human or an LLM. It has increased the likelihood that the result you get for any given term is slop rather than genuine content, for no other reason that the fact that AI content factories can produce content at a rate that Google can't keep up with.

It is not inconceivable that Google Search will become completely unusable, unless you already know exactly what you're looking for, which in most cases amounts to laziness on the part of the user to not go direct to those sources to begin with. Google Search like most of Google's products is subsidised by the revenue raised through advertising, the question remains how long are advertisers willing to pay to advertise their products to bots rather than actual people? Several years ago I remember a flurry of articles discussing the merits of online advertising and whether it was worthwhile at all, several case studies abandoned online advertising entirely and monitored their sales to see if they fluctuated at all and the did not.

Advertisers for now still believe paying to advertise through Google is worthwhile but there will be a turning point when they put that to the test and Google will suffer financially as a result. Youtube is the bellwether for this tide change, already it has stepped up efforts to demonetise channels. Around a year ago Google removed the "Play All" link from the video section on channel pages preventing users from playing all videos from a channel in a playlist, their justification was that Google doesn't deem this viewership to be "genuine interest" - a term it has vaguely defined and purposefully so in its advertising agreements. New copyright detection algorithms scan videos for audio, video, and transcribed audio for matches against copyright content and disqualifies videos from earning channel creators any revenue - the advertisers whose ads are shown still get charged for the impressions however, emphasising the purpose of this policy is not to prevent monetising copyright content but to prevent paying the content creators, Google itself is still happy to profit from it.

The inability to tell who is human online and who is not, poses an existential threat to many of the services we use every day. While it's tempting to write this off as the concern of corporations and only having implications for their profits, the wider implications on the users of those services is a deeper concern. Someone shared a post on Bluesky linking to a Guardian article espousing the virtues of a pre-internet society and how great it would be to return to it. I'm a millennial, part of a unique generation that came of age alongside the internet, we were born into a world that didn't have the world wide web application that uses the internet, witnessed its birth, the growth, and maturity of the technology. My generation has experienced the world before, and after the internet, I feel that makes me and my generation uniquely qualified to say you don't want to go back to that world.

Nostalgia is a drug, Baz Luhrmann once described it in the song 'Sunscreen' as "a way of fishing the past from the disposal, wiping it off, painting over the ugly parts and recycling it for more than it's worth" - there are a lot of things I miss from the 90s, there are a lot of memories of the pre-internet world that I hold onto, but I also remember the practicality and the pragmatism of that time.

I remember when a bank transfer took a week, when checking you balance meant calling a number and using a slow automated system or physically going to a bank or ATM to check. When the only media you consumed, games, music, movies, was that which you could afford to physically buy either legally at full price or illegally from a guy that knew a guy if you were so inclined, or that you rented paying per item to keep it for a week and return it. When most of the shops in your town all bought their goods from a local or regional wholesaler which meant they all sold the same thing at roughly the same price, and had only a limited selection.

People think of the internet and think of a browser in their computer or on their smart phone and think that's all that it is, but it reaches further into our lives. A small shop in your town in the 1980s had a limited number of suppliers that could use, relied on trade magazines and subscriptions to find new products to sell, had limited means to communicate and cooperate. Buying and selling anything from further afield was unheard of, something only large national retailers did and even then they often chose not to because demand for products didn't exist at the scale needed to justify the cost.

So much of our banking infrastructure today makes buying things easier, transferring funds and paying bills relies upon the internet. I remember my parents physically paying bills when each one arrived at the end of the month or end of the quarter there was a printed Bank Giro slip at the bottom you took to a post office and paid over the counter and that payment took a week or more to reach the company and you wouldn't know it went off your account until you got the next month's bill.

These people who tell you they want to return to a world like this are the same people that think "the internet" is social media and nothing more, they're the people that use the internet constantly but never think about it or recognise they are doing it. The ultimate irony is that the author who wrote that article and published it on the Guardian's website wouldn't have a job if they didn't have a website. Good luck trying to convince people to actually buy a newspaper again, there's a reason the pay-wall experiment failed miserably. Online media that actively uses pay-walls at this point don't do it as a means of raising revenue they do it as a means of controlling who has access and can contribute to their services. You have to pay to access Bloomberg because that pay-wall excludes anyone who can't afford to pay and limits who can interact with their content.

As for the argument that we can "keep the useful parts of the internet" and get rid of the "problematic parts" that will never work in practice for one simple reason - you already can, and you don't. Telephone banking services still exist, so do ATMs, and you can still receive paper statements. You can cancel all of your streaming services and only buy physical media, yes it's limited, it always was, if the store in town didn't have a copy of the game you wanted you didn't get it. That's how physical media works, there isn't an unlimited supply of it. You can choose to only buy local products from local retailers and not buy anything imported and not buy anything from large chains. You can cancel your internet service and live an offline life right now, but you won't, because you don't want to, and the people that write bullshit articles like that one won't either, because what they really want is for that "choice" to be imposed on others, which means its no longer a choice.

This grievance isn't limited only to nostalgia for pre-internet life in terms of being disconnected, it's something that extends to design and fashion too but again it bears repeating, there's nothing stopping you personally from living that life if you want to. You can buy 90s fashion, and eat the foods we all ate in the 90s, buy a VHS player and an old CRT TV and fill a bookcase with 90s movies and watch that content, no-one is stopping you from doing that.

Whilst returning to a pre-internet world in an effort to escape the problem of AI slop is something that some people would arguably benefit from, the detriment of taking a step back such as this reinforces the idea that it would effectively be revisiting technology designed to only benefit a narrow definition of people worthy in society to benefit from doing so.

Worker exploitation is already a problem, people often cite the affordability of housing in the 1980s and 1970s as something they'd like to see a return to, often overlooking the fact that most people weren't paid well back then, many countries, the UK included, didn't have minimum wage laws - the minimum wage was only introduced in the UK in 1999 - the law itself was passed in 1998 but many employers protested its introduction claiming it would cause mass unemployment as they wouldn't be able to afford to pay workers the wage it demanded which was £3.60 per hour at the time. You can make various arguments about house prices in ratio to earnings etc, the fact remains people feeling nostalgic for that time period are mapping their current situations onto the past not comparing like-for-like and thinking they would be better off.

If you want to return to a pre-internet society you will sacrifice the convenience and the connectivity that it provides. At the risk of sounding like a broken record though, again, no-one is stopping you from doing that yourself already. If you don't want to pay attention to the rest of the world, fine, you don't have to, you can disconnect and live that life of isolation.

That also raises the point of social enfranchisement; as a gay man who grew up in the closet in a socially conservative town who tried to take his life multiple times, having a connection to a wider gay community online would have fundamentally changed my life. I didn't experience that connection until I was in University, I was 18 by the time the internet had reached a point where I could use it in that way, my teenage years would have been completely different if I had even a single voice that could have said to me, you're not alone, you're not the first to go through this, here's how I dealt with it, here's what you can do to make it through. The fact I am even alive today is a miracle but for many others they weren't so lucky, where I failed others succeeded and are no longer with us. The value of communication and connection that the internet provides is not something that society should give up so easily.

A post-web world that still has the internet and applications, but with websites rarely used is one potential future. I don't know what that world would look like, but I am confident it would be incredibly lonely. Some people blame the internet for people engaging less with their physical communities and the people around them, but people use the internet to engage with people that they have a genuine connection to, shared interests, and commonalities. People don't interact much with the people around them because they never had a genuine interest in each other to begin with, as someone who saw the world before and after the internet I know for a fact that many people pre-internet put up with others because they had to, because they were their only social connections, not because they actually wanted to.

We haven't become a disconnected society, we've become a society that focuses more on meaningful connections that has no time, energy, or patience to maintain the pretence of connections we don't care about. We value our spare time more because so much of our lives are spent earning money to live, that when we get a chance to live it we don't want to waste our time doing things we don't actually want to do. We have become more self-indulgent because we needed to be in order to survive. None of this is the fault of the internet, this is the fault of capitalism, and before those on the right chime in with "capitalism gave you the internet" - it did not. What you call the internet is the world wide web, a software application developed by Tim Berners Lee who worked at CERN, the European Centre for Nuclear Research, an intergovernmental organisation which was comprised of 16 countries at the time, paid for by tax, it was ultimately a project developed through public money not private, and the patent for it was made open and free to use. The web exists not because of capitalism but because of socialist ideology, that common good and cooperation benefits all and is a worthwhile endeavour, it was not pursued for the sake of profit at all costs.