< Conversational AI ended copywriter profession as we know it - vol. 2 - AI Marketing HackerBlog o marketingu a AI. Conversational AI ended copywriter profession as we know it - vol. 2 - AI Marketing Hacker

Conversational AI ended copywriter profession as we know it – vol. 2


What can conversational AI do? Write texts. Good texts.

This is a vol. 2 of an article on human copywriters vs. conversational AI tools. Read the vol. 1 here.

The battle copywriters vs. conversational AI is on. With the help of good prompts, LLMs can easily manage everyday marketing-communication texts and as repeated tests show, AI can handle academic texts. So it the hope for human copywriters lost? What are the potential problems of using conversational AI? It can for sure “write”, what can’t it do?

Personal meetings

Thank you, Cpt. Obvious, you might say. The issue is though, that the vast majority of copywriters I’ve spoken to have refused to meet in person. At best, they agreed to an introductory video call. But the rest of the work was to be done through a system of written brief, documents on an online repository, submission of the final result online. It seems that the good’ole warm human contact reduces performance. Let’s compete in performance with the computer, what can go wrong?

Being creative and having a human creative force

You can ask it to “invent” the subject of a text or even an entire project, but it will be a synthesis of knowledge with an output that sounds like a good idea or truth, but may not be. Mr. Jackson Greathouse Fall had the idea for ChatGPT to “invent” a business with a starting capital of 100 USD (you can follow his fantastic story live on Twitter), but Mr. Fall had the idea, not ChatGPT.

Managing other platforms (for now)

Writing a text and submitting it in a document is one thing, its journey to going live, whether on social media, websites, chatbot backends, etc., is not necessarily lengthy, but typically manual and full of “clicks,” far from the idea of fun, creative work. Online tools working with platform APIs can solve this today, but it’s not 100% and truly automated, error-free, and without glitches during updates according to my personal experience, only very expensive platforms like FalconSocial can do it. Conversational AI is not there yet. Let’s see how long this paragraph lasts.

Tackling controversial, explosive topics

The “woke” agenda affects artificial intelligence very intensely. A machine evaluating the cost-benefit ratio without concern for ethical issues will, of course, conclude that slave labor is an excellent way to get work done with minimal expenses, for example. For years, humanity has been trying to correct or censor content. Much has been written about ChatGPT avoiding edge topics, even according to various languages or regions. People like Jordan Peterson spend hours trying to uncover censorship on this AI and more hours getting upset about it.

Creating unique content.

Digital tools inherently offer easily replicable output. Same prompts, same results. Slightly different prompts, the same results again. Texts generated by them on a given topic are practically identical. Most people write prompts in English and basically the same way. Of course, there is a huge bundle of texts that already are and always have been almost identical, regardless of the author: product descriptions of a certain type of goods, texts related to tourism, restaurant menus, washing machine instructions… However, this means that the world from now on will be flooded with a wave of texts, from which their authors will rejoice, but will be desperately non-unique, replicable, and thus SEO-problematic. Especially for people with low functional literacy (in some corners of the world, this may be the majority of the population), AI-infused texts are simply phenomenal at first glance.

If the input for ChatGPTv4 is:

“Make something with da yuh pig living next doa di mess yuh leaving behind round di garbage area makes fi wi neighbourhood unliveable I need a sign”

and the output is:

then every conversational AI user is a copywriter with outputs whose quality galactically exceeds the quality of the input, but they are very similar to each other regardless of location, education, language…

Copywriters refuse being human

As you can see, conversational AI has many limitations. The problem, however, is that many of these limitations are not offered by copywriters. They fundamentally refuse to offer them. The “junior” ones, unsurprisingly, mainly women, at least offered social media management, SEO adjustments directly in the CMS website, or email management. But most writers required exclusively writing, and in peace, without interruption, i.e., without human contact. Work always from home, focusing on copywriting-only in silence. They have reduced themselves into a copywriting tool – and this will lose to copywriting tool, inevitably.