The Future of Copywriting: Can a Robot Compete With Humans
The Future of Copywriting: Can a Robot Compete With Humans : A robot shoots TV commercials, agencies replace creative directors with algorithms, neural networks write articles. Will technology be able to replace human competence in the future?
Let’s take a closer look at the current capabilities of robo-writers and their promising future. The future of copywriting: can a robot compete with humans?
In the last few years, there has been a significant breakthrough in the field of text generation. The best robot writers are like enthusiastic beginner copywriters – their powers are inexhaustible and they produce creative, but not perfect work.
Can a machine really perform creative tasks?
When it comes to competing with machine learning, some argue that creativity is the only advantage humans have. Though AI agents have improved their ability to diagnose fractures and control drones, their creative abilities remain limited. However, robo-writers are already working in some newsrooms.
In addition, they are beginning to be used for vFairs virtual event solutions. So far, they have mostly been doing reviews of high school sports, writing about the weather, and analyzing the stock market, producing fairly standard texts.
Dynamic and data-driven creatives have long been used in the advertising industry to speed up, scale, and automate the creative process. Many different ad banner combinations are created and tested automatically, including images, titles, descriptions, and other elements.
The GPT toolkit was developed based on a massive amount of data. Millions of documents on the open web, including Wikipedia, writing services like Write my paper, and self-published books, were used for the GPT-3.
Based on this text, the model builds a vocabulary of 50 thousand words. Its scales predict the next word based on the previous ones, allowing it to go beyond basic memorization and improve meta-learning. It will need a hint and will be led by context-rich text samples.
The GPT-3 can quickly generate a logical snippet of text from a hint like “Car insurance is…”, as well as generate sequential text based on an action movie and comic book plots. In this case, the model’s texts are not simple plagiarism. It learns to predict words based on what has already written by others, which makes its writing predictable. In reality, however, this is what most writers do.
The largest GPT-3 setup includes 175 billion parameters. They can be compared to a little leverage that models need to adjust, based on examples of human texts, to predict the next word in a sentence. If you give GPT-3 hints like “A long time ago…” or “This is not me…”, the model will produce believable text, even a little creepy at times.
This is the main difference between the new generation of robot writers: they are flexible models, bordering on the field of general artificial intelligence (Artificial General Intelligence – AGI).
It is not trained in any particular discipline, but it is capable of learning. The GPT-3 appears to be capable of a wide variety of linguistic activities, including translation and communication. Using such a tool to create creative hints that will then be handled by humans will bring many benefits to advertising copywriting.
Predictions for the future
GPT-3 also has some limitations. It forgets the original context after around 800 words and, so it is better to use it for short texts. This model has a short attention span, just like humans.
As a result, those who have had the opportunity to learn more about the model are less impressed with its capabilities. “As you read more and more examples of GPT-3, particularly long texts,” one tester said, “your initial excitement inevitably wanes. The GPT-3 has a tendency to lose track of thought when working with long fragments.”
Ad copywriting, on the other hand, is created by trial and error and often needs inspiration. And it’s possible that in the future this inspiration will increasingly come from the robot.
The Future of Copywriting: Can a Robot Compete With Humans Written by Helen Wilson
Helen Wilson is a professional content writer. Her main spheres of specialization are AI and Business. She also studies topics about psychology and health, and also she provides help on how to write a book report.