ChatGPT: is the neural network the death of copywriting?

In an age when artificial intelligence is being introduced into every sphere of our lives, new questions and challenges appear. One of them concerns the copywriting profession and is linked to the arrival of a neural network called ChatGPT. The question sounds simple: «Is ChatGPT a threat to copywriters?» Can it replace human work and become the new standard in this field? 

In this article we try to make sense of that difficult question by exploring the capabilities and limitations of ChatGPT and comparing them with what professional copywriters can do. Our goal — is to give a well-rounded understanding of the issue and help you form your own opinion.

Table of contents:

• What is ChatGPT?
• ChatGPT versus copywriters: comparing capabilities
• The capabilities and limitations of ChatGPT
• ChatGPT as a tool, not a replacement

What is ChatGPT?

ChatGPT, or Generative Pre-training Transformer — is an advanced artificial intelligence model developed by OpenAI. 

ChatGPT is trained on a huge volume of text data — from scientific papers to internet publications — and then uses that information to create new content. During training the model learns the structure of language, grammar, vocabulary and even some aspects of general knowledge and culture. That lets it generate text that sounds natural and coherent. 

If you ask ChatGPT to write an article about climate change, for example, it will generate a text based on what it «knows»  about the topic from its earlier training.

That makes it a useful tool in many fields, including copywriting. ChatGPT can help create a content plan, write articles, notes, hashtags and even product descriptions, based on the data entered by the user.

ChatGPT works on machine learning algorithms. Here is how they operate:

1. Tokenisation. The input text is broken into separate «tokens» (words or parts of words). That helps the model process the text piece by piece and simplifies the training process.
2. Transformer processing. The tokens are processed by a transformer, which is the core of the model. The transformer uses attention mechanisms to determine the importance of each token in the context of the others. That lets the model understand the connections between words and phrases in the text.
3. Text generation. Based on the training data and the transformer's attention, the model generates the next token in the sequence. The process repeats until the whole text is generated or the length limit is reached.
4. Decoding. The generated tokens are then decoded back into text.

It is important to note that although ChatGPT can generate convincing text, the network does not understand it the way a human does. It has no views, feelings or opinions of its own. Everything it generates is based on the data it was trained on.

ChatGPT versus copywriters: comparing capabilities

In one experiment run on Tproger, a popular online resource for programmers, a copywriter and ChatGPT were asked to write texts on the same topic — «Why Python is popular». Both participants were given the same article structure and had to follow it while writing. 

The results were rather interesting. 

ChatGPT coped with the task, but its text was somewhat abstract and repetitive. When comparing Python with other programming languages, for example, ChatGPT kept highlighting the same advantages — adaptability, scalability, code readability and so on.

The copywriter, on the other hand, although writing a text similar to the content generated by ChatGPT, was able to bring in their own uniqueness and creative approach. The copywriter took the context into account and wrote a deeper, more informative text.

The capabilities and limitations of ChatGPT

ChatGPT has a wide range of capabilities that make it a useful tool in copywriting. It can generate texts on various topics, help build the structure of articles, generate content ideas and even help edit texts. Like any technology, however, ChatGPT has its limitations.

What ChatGPT can do:

• Generate text. ChatGPT can generate text based on the user's input. This is used to create articles, blogs, social media posts and other kinds of content. AI is especially useful for companies that publish content regularly and are looking for ways to increase its volume without losing quality.
• Help build article structures. ChatGPT can help create a structure for an article, suggesting sections and subsections based on the topic.
• Generate content ideas. AI can suggest content ideas based on a topic or a set of keywords.
• Help edit texts. The network can edit text, suggesting synonyms, correcting grammatical errors and improving the writing style.
• Automate routine tasks. Some tasks, such as writing product descriptions or creating social media content, can be routine and time-consuming. ChatGPT can automate them, freeing the copywriter's time for more complex, creative work.

And the main advantage of ChatGPT — is that it can generate large volumes of text in a very short time, 24/7.

The limitations of ChatGPT:

• Understanding context. Although ChatGPT can generate text, it is sometimes unable to understand the context or the meaning of the input. That can lead to text that does not match the user's expectations.
• Originality and creativity. Unlike human copywriting, ChatGPT is not capable of creating genuinely original, creative content. In some cases it generates content that is very abstract, and sentences occasionally repeat themselves.
• Accuracy of information. ChatGPT creates text content on the basis of the data it was trained on. As a result, some of the information it generates may be out of date or inaccurate.

ChatGPT as a tool, not a replacement

ChatGPT can generate large volumes of text quickly and never gets tired, while copywriters understand context, emotion and the nuances of language better. It is the living human being who can determine which tone and style suit a particular audience or situation best. Copywriters can also draw on their life experience and creative ideas, which AI cannot yet fully reproduce.

For all its capabilities, ChatGPT is not a replacement for copywriters. It should instead be seen as a tool that can help copywriters improve their work and become more effective. 

Predicting the future is always difficult, but it is fair to assume that ChatGPT and similar technologies will play an ever greater role in copywriting. They can become useful tools for copywriters, helping them create content faster and more effectively. It is unlikely, however, that they will fully replace real copywriters, at least in the near future.