We're talking about ChatGPT and the benefits it can bring you. Coverage keeps piling up around the software from US company OpenAI, and the whole world is captivated. But what exactly is this ChatGPT, seemingly shaking entire professions to their core?
The End of Art? A Personal Preface
One of those professions is mine — copywriter. I think I've been one since I was a small child. Even my primary school teacher praised my stories, some essays were printed and copied for younger classes, and writing free-form and academic texts came easily to me throughout university. And if playing with words wasn't already a gift (my mother's words, not mine!), it culminated in something else entirely when I registered as self-employed. 'Registering a small business?' asked the woman at the chamber of commerce, quite baffled. 'No, we'll have to do this differently. Your activity is classified as artistic.' And there it was — I'm an artist! Ever since, I mentally throw an imaginary silk scarf over my shoulder, put on my glasses, and consider Goethe and Rilke my colleagues. That's how things stood until this assignment came in: could I write something about AI content creation, ChatGPT specifically? The nerve! As if it isn't enough hearing about it everywhere already — what it can do, how effortless writing has become, wonderful, wonderful, wonderful! This artist's temper is easily provoked! But fine, as we say up here in the north: what must be done, must be done. Into the unknown, facing the enemy (that's what we artists call this ChatGPT — we are, after all, sensitive, emotional souls). Time to start researching. But coffee first!
Before I present all this in detail, let's take a step back. To understand ChatGPT, you first need to grasp the technical background: artificial intelligence (AI). Never heard of it? Surely you have. Never experienced it? Nonsense — you just spoke to Siri or Alexa to ask them what ChatGPT is. So there you go, you're already well acquainted with artificial intelligence. Because technical applications talking to you and understanding you is the whole point of this branch of computer science: AI aims to transfer human thinking and learning onto machines. And it's meant to do it all — perceive sensory input, store and process information, solve problems — the list seems endless. How this is possible could be described here in exhaustive detail. A small tip: there's a very informative Wikipedia entry on it (which I obviously didn't read during my research — I am a professional, after all). But let me sum it up like this: AI is based on a huge amount of data, processed purposefully by programmers and made available. That's exactly what OpenAI did with its ChatGPT software — just with even more input, data and sources. Technically, the chatbot can be defined as a text-based dialogue system. In other words: you enter a question or ask the system to write a text on a given topic, and get a result within seconds. ChatGPT's ability to deliver this is built on machine learning, a process where knowledge is artificially generated from experience. What makes GPT special is also its training through self-supervised learning and the vast amounts of data drawn from online forms, social media, books and much more. Now you're getting sceptical, because we all know that for every 100 pieces of correct information, roughly three times as much nonsense is running amok on the internet? You're not alone in that — but OpenAI has planned for this too: unlike its predecessors, ChatGPT was fitted with safeguards designed to avoid false or harmful answers. Now you're thinking 'incredible!' — and yes, that's roughly how the 100 million people who've adopted the software as users since November 2022 feel too. I can only repeat myself: wonderful, wonderful, wonderful! But I haven't told you yet what else ChatGPT can actually do. Honestly, I'd rather not — I might end up impressed myself and betray my own craft! But now I really want to know more.
The End of Art? A Personal Preface
One of those professions is mine — copywriter. I think I've been one since I was a small child. Even my primary school teacher praised my stories, some essays were printed and copied for younger classes, and writing free-form and academic texts came easily to me throughout university. And if playing with words wasn't already a gift (my mother's words, not mine!), it culminated in something else entirely when I registered as self-employed. 'Registering a small business?' asked the woman at the chamber of commerce, quite baffled. 'No, we'll have to do this differently. Your activity is classified as artistic.' And there it was — I'm an artist! Ever since, I mentally throw an imaginary silk scarf over my shoulder, put on my glasses, and consider Goethe and Rilke my colleagues. That's how things stood until this assignment came in: could I write something about AI content creation, ChatGPT specifically? The nerve! As if it isn't enough hearing about it everywhere already — what it can do, how effortless writing has become, wonderful, wonderful, wonderful! This artist's temper is easily provoked! But fine, as we say up here in the north: what must be done, must be done. Into the unknown, facing the enemy (that's what we artists call this ChatGPT — we are, after all, sensitive, emotional souls). Time to start researching. But coffee first!
Stumbling Block or Springboard — So What Do We Do With You, ChatGPT?
Oh, it would be so nice! But the moment you get to the world's biggest search engine, the romanticising of AI comes to an abrupt end, and you quickly wake up from the dream of instant content creation (do you hear that? Yes, that's my smug laugh — mind you, from behind coffee cup number 3): Google's slogan still holds: content is king! Because the IT giant's business model rests on one simple goal — users should get the most relevant result for their search query, one that answers their question or completes their search. It's no accident SEO experts have been battling each other for years over filling websites with the best content, including the right keywords — and doing it with genuine sense and purpose. Google recognises that effort and rewards SEO content and SEA ads with a quality factor and, ultimately, a better ranking. Can you outsmart Google with super-duper AI-generated content? Wrong. The search engine caught on to that trick long ago and is pushing back — not least because ChatGPT in particular could threaten its whole success model. After all, who still needs to search the World Wide Web when they can just ask the little helper instead? So Google created SpamBrain, an application that identifies signals and patterns of AI content and flags it as spam under its guidelines — and, of course, is working flat out on its own chatbot. Back in the schoolyard we'd have called that behaviour sulking (okay, fine, we still say it today too — you may have already thought it about me given all my grumbling!). In the grown-up business world, it was initially also an attempt to contain AI's power. But at the latest since its triumph last year, even Google has recognised that a stubborn rejection of AI content isn't forward-looking, and walked things back slightly in its early-February guidance on AI-generated content in Google Search. Google's ranking system still rewards original content distinguished by E-E-A-T. No, don't worry, reading that acronym isn't going to make you hungry like it does me — it has nothing to do with that! E-E-A-T stands for the quality signals content is measured against: Experience, Expertise, Authoritativeness and Trustworthiness. And unlike before, it's now stated that what matters is the quality of the content, not how it was produced. So not every automation or AI generation is automatically classified as spam. 'Appropriate use of AI or automation does not violate our policies,' the guidance states — though everyone instantly stumbles over what exactly counts as 'appropriate'. That word alone smells like a catch! And straight into the next tangle we go. So you're allowed to use AI, just not to manipulate search result rankings. Fair enough — but since chatbots like ChatGPT don't cite sources for their answers, Google struggles to detect E-E-A-T signals, and penalises the content as a result. It also stresses that SEO texts need to be enriched with genuine human experience. So there it is: not everything is quite so rosy in practice.
Time to take a breath — let's do exactly that. Because to apply all of this properly in the online world, there's really a lot to examine. But should I dive into the world of ChatGPT and face the enemy — and, in the end, Google too? 'My working methods would stagnate if I don't stay up to date,' the little devil on my shoulder whispers to me. The angel, deeply tied to my emotional side, left the building hours ago. But the devil has a point — I can't keep refusing this any longer, so I fire up ChatGPT once more.
It's evening by now. My coffee cup has been replaced by a glass of white wine. The younger generation would call it 'treating yourself'; the self-discovery gurus would call it 'me time'. Either way — I've complained a lot today, but I've also learned a lot. The most important realisation: ChatGPT doesn't have to be my enemy. More than that, I can turn it into my little helper. Because the moment I opened it, one thing became clear: if I don't talk to it and tell it exactly what I want and what it should do, nothing happens. Really, nothing at all. Give it vague instructions, and nothing useful comes out. And once again I realised: the chatbot may be intelligent, but without the person sitting at the laptop or phone, that intelligence never gets tapped. If I just type 'write me a text', I get one — but nothing I could actually use. I have to be precise and know exactly what goals I want to achieve with its help. So the AI doesn't steer me — I steer it. (Well, it's not going to get any more poetic than that!) Does that reassure me? Yes, a little. Because I also had to admit the style of these texts isn't quite polished yet. Repeated words aren't uncommon, and I often had to go back in, replace things, rephrase things. A wave of relief washed over me shortly after: phew, I can put off that retraining application a little longer after all — not least because I don't necessarily have to use ChatGPT's texts, but I can draw input from them. A new synonym here, a fresh angle for a description there — all inspiration I'm happy to take. Okay, fine, I'm warming up to it. But keep this in mind: don't overdo it! Google is watching, keeping an eye on me, and it's even worse than my little devil. Who, by the way, hadn't been entirely idle and whispered to me that the AI Text Classifier tool can help you play it a bit safer: you can enter texts you've written with AI assistance and check whether it's immediately obvious that parts of them are AI-based. Maybe that gets you ahead of Google's guidelines?
You can never be entirely sure, but it does make you feel a bit safer. Well then — cheers to that!
