Working With AI — My Job and My Hobby

I work with AI every day. Partly because it’s my job, and partly because I genuinely enjoy it.

At L’Oréal, AI is already deeply connected with how the company approaches technology, beauty, marketing and digital development. It’s one of the reasons I find my current work so interesting: I get to see and use technology that goes far beyond simply opening ChatGPT and asking it to write something.

In my work, I use and work around tools and technologies including ChatGPT, Claude, ModiFace, Beauty Genius, Noli, CREAITECH, Salesforce AI and WPP-NVIDIA solutions, alongside predictive models, augmented-reality try-ons and diagnostic technologies. AI also increasingly intersects with content, localisation, search and the multilingual digital work I do across different markets.

My background is in SEO and digital marketing, so experimenting with new technology has always been part of what I do. Search itself has changed constantly throughout my career, and AI is now changing it again.

But my interest in AI definitely doesn’t end when I finish work.

Then I go home and build things

I use AI extensively in my spare time, and some of my experiments have become proper projects.

One of them is my own little "Jarvis".

I took an old laptop and started turning it into a personal AI assistant. ChatGPT essentially functions as its brain, but the interesting part for me is connecting the language model with other components: speech recognition, memory, tools, APIs and voice.

The point wasn’t to create a commercial product. I built it because I wanted to see how far I could take the idea myself.

That’s generally how most of my AI projects begin: I wonder if I can make this work.

Sometimes the answer is yes. Sometimes the answer is eventually. Sometimes I create a spectacular technical mess and learn something from it.

All three are perfectly acceptable outcomes.

AI also became part of my creative hobbies

My experiments aren’t limited to work, SEO or automation.

I make music under the name Groovy Diesel Spaghetti, a project I originally started years ago, long before the current generative-AI boom. Today AI has become another instrument in that project.

I use it as part of the process of creating music, experimenting with sounds, developing ideas and producing visual material around the songs. One of my recent releases, Already Mine (Dreamcatcher Version), was created largely with AI.

For me, that doesn’t make the project less creative. It simply means my toolbox has changed.

I also use AI to create art, images and comics.

One of those experiments became Kevin the SEO Manager, my satirical comic about the absurdities of working in SEO and digital marketing. It features an increasingly unfortunate SEO manager dealing with Google, ChatGPT, Claude, SEO tools, CEOs, brand managers and the sort of wonderfully ridiculous situations anyone who’s worked in SEO for long enough will recognise.

AI gives me a way to turn ideas that previously existed only in my head into something visual.

That’s what interests me.

I’m not trying to prove that AI can replace an illustrator, musician, developer, writer or anyone else. I’m interested in what I can create with tools I didn’t have before.

Apparently, this is controversial

The slightly bizarre part is how much criticism this interest has attracted.

There’s a surprisingly hostile attitude towards AI in some professional circles, particularly towards people who openly say they enjoy using it. I’ve heard practically every variation of the argument: AI makes people lazy, AI users can’t think for themselves, AI content is automatically rubbish, AI destroys creativity, and using AI somehow makes your professional skills less legitimate.

I find some of this criticism rather odd.

Using AI is quite literally part of my paid professional work.

Outside work, building AI systems, making music, creating comics, experimenting with images and testing new technology are my personal hobbies.

I’m not entirely sure why either requires justification.

Of course AI produces rubbish. So do humans. I’ve spent enough years working with websites and content to know that mediocre content didn’t suddenly appear with ChatGPT.

The same applies to creative work. Having access to a tool doesn’t automatically produce something interesting. Give ten people exactly the same technology and they’ll create ten very different things — including, inevitably, quite a lot of rubbish.

That’s not particularly different from what happened before AI.

I don’t use AI to avoid thinking

One assumption I find particularly strange is that using AI means outsourcing your thinking.

My experience has been almost the opposite.

The more I use these systems, the more interested I become in their limitations. I test what they understand, where they fail, how context changes an answer, how different models behave and what happens when they’re connected with other tools.

Building my Jarvis is a good example. The interesting part isn’t simply having an AI that can answer questions. I already have that on my phone.

The interesting part is figuring out how the components work together.

The same applies to music and visual art. I can generate something in seconds, but getting something I actually like is an entirely different matter.

Sometimes the result is useful. Sometimes it’s beautiful. Sometimes it’s hilarious. Sometimes it’s complete garbage.

All of it teaches me something.

SEO made AI particularly interesting to me

My fascination with AI also makes perfect sense in the context of SEO.

Search professionals have always worked somewhere between humans and machines. We try to understand how information is crawled, interpreted, organised, retrieved and presented.

Now that relationship is changing.

People increasingly ask questions directly from generative systems rather than typing traditional searches and choosing a website from a list of links. Brands therefore have to think not only about traditional search visibility, but also about how their information is understood, retrieved and represented by generative systems.

That’s why I’m interested in GEO as well.

For me, GEO isn’t some fashionable replacement acronym for SEO. It’s another development in the same field I’ve worked in for years.

I don’t know exactly what search will look like five years from now. Nobody does.

That’s precisely why I’d rather experiment with the technology now than spend the next five years arguing about whether it should exist.

You don’t have to like AI

There are legitimate problems surrounding AI, including copyright, privacy, misinformation, bias and the effects of automation on employment. Those deserve serious discussion.

But serious criticism of technology is different from hostility towards somebody because they happen to use it.

You don’t have to like AI.

You don’t have to use ChatGPT or Claude. You don’t have to generate images, make AI-assisted music, automate your workflows, build your own Jarvis or spend an unreasonable amount of your spare time trying to make an old laptop do things it was never intended to do.

I happen to enjoy all of that.

I’ve spent much of my career learning new digital technologies as they appeared, from websites and search engines to smartphones, social media, analytics, automation and now artificial intelligence.

Some of today’s AI tools will disappear. Some of the things currently marketed as revolutionary will turn out to be completely pointless. Others will quietly become normal technology that everyone uses without thinking twice about the AI behind them.

I want to understand which is which.

And for me, the best way to understand technology has always been the same: use it, test it, break it, build something with it and see what happens.