AI Can Write Stories… But Is That What We Really Need?
Its been more than a week since the world read the op-ed written by GPT-3, a “autoregressive language model that uses deep learning to produce human-like text”, published in The Guardian. We all had so many questions about that op-ed; from the choice of topic, to the perspective, to the very not-innocent tone, and finally the note added by The Guardian editor saying that “GPT-3 produced eight different outputs, or essays. Each was unique, interesting and advanced a different argument. The Guardian could have just run one of the essays in its entirety. However, we chose instead to pick the best parts of each, in order to capture the different styles and registers of the AI.”
Later on, I saw similar experiments with producing stories and articles written by AI software. One of them was generated by Noon Post, a digital independent news website in Arabic. This time, a GPT-3 model, named Oppy, wrote an op-ed about immigration, a controversial topic in a region where immigration has become the norm, since its countries cannot offer anything to its citizen especially the younger generation.
The ope-ed by Oppy was originally written in English and then translated to Arabic. It puts forward the pros and cons of immigration. It mainly tries to answer two questions: what are the opportunities and dangers of immigration, and what are the best countries to immigrate to. Through answering those questions, I thought the article is more of an explanatory article more than an opinion piece. It’s just trying to be on the fence! Personally, I don’t think this is how op-ed works. I believe all what they are trying to do is avoid the controversy that could be created as a result of a “robot writing that piece” rather than a human writing it.
Another project by Cairo24, a digital news service out of Cairo، aims at producing business stories using this technology. According to Mahmoud Najm, a business researcher, Cairo 24 used automation to produce business stories that would be written in a more simplified way for the audience. In an interview with IJNET Arabic, Najm said that this way also helps in giving journalists more time to focus on producing more investigative stories. This is an example of story written by AI through automation. The story talks about an increase in the number of jobs on the jobs portal “Forasna”.
Quantity… or Quality Problem?
Many claim that AI could be used in newsrooms to write stories and articles that come from known sources, need to be updated everyday, and have an importance because they are current affairs. These stories include stock market news, oil prices, and football game results. In other words, AI could do the “boring lengthy jobs”, and leave the more creative and interesting work for humans to do. Theoretically, that makes sense. But from an audience point of view, this means adding a bulk of boring stories just because we have to have them. In this particular case, I believe AI could assist in the collection and gathering of information, but probably in this type of stories, it’s not about publishing them, but about finding new ways of storytelling to make them more interesting.
A Recycle of Ideas… FULL STOP
Writing has always been the product of ideas, experiences, and interaction. Until now, machines have been learning from past ideas, experiences and interaction. So whatever result machines are going to give us is going to be a recycle of ideas, experiences, and interaction.. In short, the path of creativity will stop there.
Take the post written by Oppy in Noon Post. It did not add anything to what we already know. It just presented opinions and counter-opinions on the topic. And most important, it was a generic op-ed; meaning that it takes the stories, struggle, and opportunities out of context. In this type of stories in specific, and talking about such a humanitarian crisis, I expect to read moving stories of leaving home and family, or happy stories of success in a new country. I did not read any of that, just because an AI model did not live through them, or did not meet someone who lived through them.
Don’t get me wrong. I am not against AI. I believe there is a great opportunity there when it comes to helping journalists in their work. But probably ending up with taking stories from robots is not the best employment of these technologies.
I would like to end this article with a short story I read about this whole scenario of reading articles written by AI. This excerpt was used by Thomas Baekdal in his article “I’m not impressed by the Guardian’s OpenAI GPT-3 article”
In 1999, Alan Cooper wrote an excellent book called “The Inmates Are Running the Asylum: Why High Tech Products Drive Us Crazy and How to Restore the Sanity”, where he said this:
It’s like the fellow who leads a huge bear on a chain into the town square and, for a small donation, will make the bear dance. The townspeople gather to see the wondrous sight as the massive, lumbering beast shambles and shuffles from paw to paw. The bear is really a terrible dancer, and the wonder isn’t that the bear dances well but that the bear dances at all.
