Not AI Again

This is not another rant about AI, I promise. AI is everywhere whether you think Skynet is around the corner or it’s the best thing since cave paintings. The subject is unavoidable. At a literary evening I went to authors were asked about their thoughts on AI in the writing world. They were diplomatic because publishers are a mixed bag when it comes to their stance. No one wants their work stolen, although every author’s work had been, while publishers think that AI can automate the more time-consuming tasks (which it can’t at present).

Already we have had the madness of Grok, every AI company being shady, unsafe and immoral while appearing to have not only gotten away with it, but to have been patted on the back from investors for doing so. Stealing peoples work is de-regur and we have been rendered powerless in the face of it. Indeed, the prevailing zeitgeist seems to be that if you are rich enough, either as a corporation or as an individual then no laws or standards of morality, decency or humanity apply to you. Conspiracy theorists are saying “I told you so” out loud.

Add to this the financial disaster of AI companies burning money quicker than a lottery winner in Vegas, promising impossible infrastructure build outs that, (unlike the dot com bubble which laid down fibre cable that has subsequently been of utility), will leave incomplete, empty and obsolescent data centres while simultaneously destroying the environment, diverting essential infrastructure investment away from people, while the rich get richer while the poor get poisoned and even more poor, then we have a bleak picture. (And a very long sentence – sorry)

It is easy to present this as a binary distinction, AI bad, not AI good. Let’s be honest a lot of AI is bad. There is a big however, firstly AI is not going away, we are stuck with it, like the spell checker, it is going to be integrated into the digital systems we use every day. And, like the spell checker it is used by lazy people to avoid putting in the effort and doing the work, but like grammar and spell checkers, ultimately it is the human that is going to produce the highest quality of work. What works best, as I type this in Word, is a human using the tools to increase quality not quantity. One problem with AI is akin to having only hammers in your toolbox, everything becomes a nail. What makes it worse is the promises of the tech bros, such as “intelligence” among other things.

AI is a token-based prediction machine which programmers have designed to be, in various degrees, sycophantic. It hallucinates because it has been coded to please and give an answer whatever the situation. AI says a lot about the tech people who code it. Sure, it can say “I’m sorry Dave, I’m afraid I can’t do that” but it can’t say “I don’t know”. This people pleasing leading to the bypassing of the guardrails and safety’s programmed in is a mirror of what happens when humans don’t enforce their own boundaries. That is to say it all goes wrong and psychosis may follow.

So, let’s call it what it is, a shit show. We are, understandably, missing that AI is a tool with use cases. I am not an AI bro, I am not advocating anything, instead I am going to explore the source of my personal ambivalence. I am completely against AI taking jobs from people, the thefts of work and the ongoing environmental destruction. That is all unacceptable. However, (that word again), Pandora is out the box, we have a tool and it has use cases. If we look not even that deeply at technological progress it has always met resistance, writing was opposed, painting style changes violently resisted, musical styles, the computer, all have, and quite rightly been questioned. Scrutiny is wise and necessary, boundaries are essential and the prioritisation of human welfare should always be paramount. Those are the essential basics, so why am I ambivalent?

Cards on the table, I do like tech, while I may not be an early adopter in the purest sense, I have been using computers since my ZX spectrum and my first PC was a parts bin based 286 using dos with qdos, 5 ½ inch disk based MS Word, a tractor printer and 16 colour VGA display. It was basic but it did everything I need it to do. I have used windows, mac os and linux, although were it not for adobe I would be on linux because it has gotten so good since I first used ubuntu back in 2006 ish. I am not technophobic, although I loathe voice commands, literally a vitriolic hatred of them.

Among other things I am dyslexic and that 286 with word and the most basic of spell checkers was an absolute revelation. Access to a computer, I feel, was a big part of why I was able to succeed at university. I had friends who proofread for me who were hugely helpful in me presenting my best at my best because they picked up things computers don’t. But being able to write and manipulate text, write plans and change them without constant rewrites, crossings out and the like, literally changed the game. I still lean on my word processor software heavily. What has this got to do with AI, for one AI can be used to strict proofread putting what it sees as mistakes in bold – no suggestions, no rewrites, but highlights. I obviously use the spell and grammar checkers – especially for commas. And I know there are still errors, it’s hard to proofread your own work.

I obviously use AI for image generation. I do this because, one stock images are not always perfectly fitting for my piece, and second, more importantly, I don’t want to spend more time looking for the image, which is literally only there because that is how the internet works, than the actual writing of the words. For me, who is a very text based creative, images are a headache. If I was super wealthy and could have a graphic designer on retainer, absolutely, I don’t. I am like most people a one-person team.

My other use of assistive technology and increasingly AI is helping me research. Between more and more going digital and my limited ability to access a library (especially the nearest University one), online tools do some heavy lifting for me. For example, I had a business idea, it had legal constraints, so I fired up a project management tool and set about deep diving it. I thought, let’s see what AI can do. It did a lot. I used it for preliminary research, to find the laws, the forms, web sites and resources to download that would help me go from idea to launch. I also used it to create some templates for certain types of submission that have a definite format. All of which I read and checked. None of it was intended to replace writing something like a funding bid myself, it was to have an idea of what a funding bid would look like for my idea and for me to have something to compare to any other examples I could find. Between myself and the resources the AI turned up I was able to learn way quicker than had I had to do that initial resource trawl through Google myself.

Thing is I have done this before, I have launched a business, written a business plan, investigated legal constraints and requirements, it was hundreds of hours of work in total. Comparing that to doing the same assisted with AI, I got to a similar point of understanding in less than 20 hours of actual work. This new idea can’t go anywhere further because of the most complicated set of circumstances beyond my control and has to be parked likely to go no further. I don’t have a hundred or more hours to invest on the off chance an idea will germinate. What AI has done is allow me to massively compress the discovery phase, which in turn allows me to do more discovery.

There is another, important aspect, I mentioned my dyslexia, I am also disabled, have chronic life limiting conditions. The adhd and autism were also there, but undiagnosed, leaving me stumbling through life baffled as to what was going on. The use of tools, like AI which I use to assist making my poem videos and doing their voiceovers, allows me to overcome the limitations that otherwise present absolute barriers to my participation in creativity. The internet and the digital world give me a life that didn’t exist when I was growing up and I am grateful for what is available to me now. I don’t believe AI will save the world, only people can do that.

I know this has been long, but I want to end with a more personal story about the digital world and AI. Back in 2017 I had a stroke. It wasn’t treated; in fact, I was sent home unable to move my left side or talk properly because I was “putting it on”. A mini saga of medical negligence that would end about a month later with me self-discharging. Because the Dr’s would not say I had a stroke, I got no follow up, nothing. My GP did their best for the medications. But I was on my own. At that immediate time the internet was essential in finding resources for my recovery. However, it was very difficult and while I have made a great recovery there have been nagging questions because I am not who I was before. On this journey knowing I am AUDHD has been tremendous for self-acceptance and management. However, (again), there were questions. I have researched the type of stroke before; I didn’t need that. The problem was accessing detailed information on the exact type of stroke I had. That was in textbooks and on Dr’s sites or journals I don’t have access to. So, what about the AI, if they have scraped the internet then maybe they had the detail I am looking for.

Skipping to the end of the story, a couple of hours of AI guided research was everything I could have hoped for and more. I explored links and correlations that I could neither prove nor disprove in the literature I had found up to that point to discover whether they were actual, plausible or not possible. I was able to more accurately pin down what areas of my brain were affected, by correlating presenting symptoms to brain area function. Like using a dictionary where you need the first letter or more, medical discovery usually relies on you knowing the correct terminology to find what you are looking for. Importantly not only did I discover what the lasting deficits could be, I found resources to explore that give ways to offset those. The emotional and cognitive aspects, for me, have been more severe and long lasting than the immediate physical impact. I understand much more about why I function, and dysfunction, the way that I do. I even discovered that the effects of my strokes are the same as dysfunctions found in adhd and autism.

So, while I am utterly dismayed and appalled what is happening around AI, and its deplorable deployment, weaponisation and nefarious utilisation, Pandora’s box does have genuine utility inside. It is a genuinely helpful tool. I don’t want an agent or that shizzle, and limited as I am, no humanoid robot either, instead the income to get the pain meds I need and a real human cleaner to help out would be much better uses of a fraction of the money wasted on AI, for me and others like me. People should come first.

Feel free to share your thoughts in the comments, I really enjoy the different takes on AI and how it is unfolding. The online content is rivalled only by “the files” in pulling me down the rabbit hole at present. Thank you for reading this far, and until next time; keep on keeping on.