If you’ve read any of my articles or blogs where I mention Artificial Intelligence (AI) you’ll be aware of my stance. As with all things tech, AI is a force for both good and for ill. In the right hands and with the best intentions, it’s a wonderful tool. In the wrong hands, it’s a blunt instrument and can even be a weapon.
It’s the same with social media. I joined Twitter in 2009 IIRC and immediately loved the space. It was great for personal interactions, and also for marketing business. Now though, I find “X” as it has been called, an unpleasant place. Too many positives have been lost in favour of what is supposed to be “free speech”. I actually don’t think people should have the right to say anything they want when that extends to hatred.
But I digress…
I came here to share my personal policy on generative AI, particularly for this personal and professional blog. It is this…
Gen AI Images
I now often use generative AI to create imagery for my blog posts. They’re not always perfect images and, truth be known, the most discerning viewers would probably call them “AI slop” at times. This is a term for poor quality gen AI images, and these are often characterised by poor spelling, or illogical things such as people with seven fingers – yes, hands are challenging for humans to draw, and now the machines are finding it tricky too.
My justification of using AI generated imagery is that I am currently limited in my use of “traditional” computer tools – I used to absolutely adore Macromedia Fireworks because it integrated so seamlessly with Dreamweaver. I was a web designer back in the day (1998-2005) and still do website design on occasions, but when Adobe took over and switched of the licensing servers after about ten years of CS3, I relied on the official codes to keep my desktop software running. I have since lost those codes and now have no Adobe Fireworks CS3.
Replacing Fireworks with Photoshop has been on my mind for a few years. I designed a website for a local custom car workshop about ten years ago, But I was unfamiliar with Photoshop, slow, and not using it as intuitively as I did Fireworks. I had GIMP and have downloaded Inkscape, but I#m so heavily involved in content and digital marketing that design and illustration has taken an ever further back seat,
So I’ve mostly been using free stock imagery and now I’m asking Copilot to generate “quick and dirty” images for my blog posts.
That’s where I stand today with AI generated images.
Generative AI Images Case Study
Working for a digital marketing agency one of the issues I encountered with a client’s ecommerce website was that their product images were very uninspiring. Their products, all ingredients for the food manufacturing industry, were discs of white powder, brown powder, yellow powder, off-white powder etc. Taking the “sell the sizzle not the steak” mantra that’s been a bedrock of marketing for years, I suggested that the raw ingredient and the use cases might better illustrate the lifeless products. The client said yes.
Now the big old traditional issue of taking the next step and arranging a photoshoot of all the hundreds of ingredients and their use cases would have been a prohibitively expensive project in itself. You’d have had to source all the raw materials (Fruits, nuts, leaves, plants, etc.) plus the supplied form (Powders, tins etc.) and the final products (Sauces being poured, chocolate coatings, glazes and so on) which would have been a mammoth undertaking. It was doable, but what using generative AI?
One of the designers at the agency then set about using Adobe Firefly to produce fresh new imagery for every product in the client’s catalogue and, hey presto, they had a wonderfully illustrated new ecommerce website.
Immediately they had “kerb appeal” and the website looked ten times better than when it simply displayed different shades of powders. Combined with some additional web design, layout changes, technical SEO, content updates, and on-page SEO, the website’s ranking, user experience, and engagement all improved considerably.
All credit goes to the designer for their incredible work.
Gen AI Text
I do use generative AI behind some of my blog posts here. This particular post does not – this is all 100% human generated and from the heart and head. Totally HI.
However, for some of the more in-depth and technical blog posts, i’ve used AI for research purposes. Generative AI does help with thrashing out a framework for a piece. I usually know what I want to write and the information and messages that I want to convey. However, running that through Perplexity, Gemini, Claude or ChatGPT “doesn’t hurt”.
Using generative AI for text, it can help affirm my line of enquiry. Seeing the same concepts and data is very validating. But gen AI can also generate some additional points too that I may have missed. Yes, it is inspiring. I also use it as a fact checker. Once I’ve written something, have I got it right? Is it factually correct? Is everything in the right order, did I miss a vital point?
Working on your own blogs and articles without the luxury of an editor, generative AI is so helpful.
So yes, I generative AI to help me with text but this is all ultimately 100% human inspired and human created. AI is a most helpful and able assistant.
Generative AI is Here to Stay
Going back to where I started in this piece, gen AI really took off when ChatGPT was unveiled back in November 2022 and my immediate reaction was of concern. I could see the opportunities but also the threats too. As a writer of long standing (Published since I was 11 years old) I’ve made a career out of putting words onto paper and now screens. It’s put food on the table, whether I’ve written pages, articles, blog posts, ads or social posts. So when people with no skills, no talent, no history, and no experience of writing could suddenly produce reams of copy for websites, that set the inner alarm bells a-ringing.
Some of the worst offenders were publicly touting to others that you could use AI to produce hundreds of pages of content in mere minutes. That was a threat to my livelihood. As a long-standing defender of quality on the world wide web that was offensive too. And so I relished every official mention of caution with AI generated content. But Google flip-flopped, first railing against, and then nuancing its stance on AI content. Google had fought against “thin content” and spammy links, and now there was this new monster in the room.
At the same time I watched the webspam reports every year. The official 2020 report from Google said they’d found 40 billion spam pages on the web, every single day! Then after that it became so much that they even gave up reporting the true numbers – by 2021 we were on treated to knowing that they found X times as many websites as the year before – my estimations are that there are now potentially trillions of pages of web spam every day.
My point here is that webspam is a massive problem and AI generated content is a big part of that problem. And I waited and I watched as Google… did nothing about it.
So this all started in November 2022. It wasn’t until the Google March 2024 Core Update that they finally made a big move to combat AI content spam. As well as the usual quality and helpful content updates, Google introduced new spam policies, and the most relevant of their new terms of services was “scaled content abuse”. This was music to my ears. I wrote about the March algorithm update and even called it “the purge”. This was defending against thin content and unhelpful content all over again.
Finally
And that is the background to my opinions and position on generative AI images and text.
I do apologise to any designers that I have chosen to use Copilot to create supporting images, infact, being a designer myself, I’m doing myself out of work. But it’s a small aspect of posting for pleasure and public service. If circumstances change I may have to as well, but for the meantime, if you need me for any projects, just call me on 01252 692 765 for any content, copywriting, SEO, PPC, and other digital marketing services.