top of page

I’m a sci-fi writer. Here’s my opinion about ChatGPT

This month’s post deviates slightly from the content I usually make. Today I will talk about a surreal real science. One that looks like science fiction, but it’s already out there, within our reach and spiking controversies: ChatGPT.


Why would you care about my opinion? Well, for one, I wrote about this technology in 2018 in my sci-fi anthology Artificial Rebellion (and now the titular story is sadly outdated because of ChatGPT).


Snippet from the cover of Artificial Rebellion

Outside the creative and educational industries, ChatGPT is considered more a tool than a threat. The thing is, I’m in both the creative and educational sectors! I'm a sci-fi writer and a university professor. So I understand people’s concerns about generative AIs and the uproar they caused in the writing community.


This is not a comprehensive essay on the topic. It's my reaction after observing speculative-fiction magazines having to deal with loads of spam submissions of AI-written stories.



A tool or a threat?


My love for science fiction comes from my perception of these stories as thought experiments on humanity and society. I see us, sci-fi writers, as oracles assessing possibilities for our collective future, be they bad or good.


Artificial intelligence is a technology full of potential. Maybe because I am inexplicably optimistic, I believe it will bring more constructive experiences than harmful ones. So, the short answer to the question posed is it's a tool.


It is true, however, that we need to bring awareness and learn how to deal with the problems it creates. To train these AIs, they must be exposed to a large data set. That's when ethics can be easily breached. One person alone cannot build a big enough database to train generative AIs; they must harvest data from somewhere.


Data mining is one of the most urgent ethical issues to be addressed here. Most people have no idea the content they produce or share on the internet could be harvested without their explicit consent. The most obvious example of violation of privacy is the image generators in which the AI has been clearly trained with copyrighted art. And, even when consent to usage is asked, very few people understand what they are consenting to. The ethics of data mining is still questionable.


People should be educated on how much their data is worth before giving it away for free.

If it is a well-informed decision, training neural networks is kinda cool. The potential for this technology is exciting, and ChatGPT is a fine example of it.



Scary AI?


It is human to be afraid of the unknown. Any novelty spikes a primal fear in us, so a varying degree of opposition is to be expected in these situations. Soon enough, it becomes part of our daily life and, interestingly, the opposite effect happens: we come to trust it too much. (If you are about my age, remember when our parents once warned us about the dangers of the internet and now believe absurd information shared on social media?)


It is natural that an AI that can write would also generate controversies. But I honestly think that, after the novelty fades, people will start to get amazingly creative with these content generators. And we’ll admire them for it.


As I mentioned, the use of ChatGPT is still seen with suspicion in the creative and educational industries, as it was not yet incorporated into the creation and teaching processes. But in the other sectors, in scientific endeavors, for example, it could help us to break barriers.


I am a writer and a professor... and also a scientist. You see, the international language of science currently is English, the native language of almost 17% of the world's population. This means that English is, at best, secondary for over 80% of the people in the world. Scientists not born in an Anglo-Saxon country must master their disciplines plus be fluent in writing in a second language to try and get their work recognized. This additional requirement could be easied with the help of AIs like ChatGPT.


Science is a collective endeavor that benefits enormously from different points of view and diverse inputs. Breaking barriers for more people to openly communicate their results is what academics should strive for.

I'm Brazilian, and Portuguese is my native language. I started writing this blog to challenge myself to improve my English skills. Although I love writing in English, it is time-consuming, the results are not immediate, and I still have to use apps like Grammarly and Google Translator to help with my prose. If ChatGPT can make this process easy for another non-native, it would be great.


So, am I using ChatGPT to write?


No, not yet. As I said, my goal is to improve my skills in English, and I don’t honestly know how to use ChatGPT to do that. It may come as a surprise, given that I’m a sci-fi writer, but I’m not well versatile with new technology (true to the Brazilian saying “In the blacksmith’s house, all knives are wooden.”) My husband, however, is quite enthusiastic about it, and I like to listen to him talk about the current and future applications. He inspires me to think deeply about these topics.



Authorbots


The idea to write about a generative program that could create whole books after being fed a great amount of literary data—as featured in Artificial Rebellion—did not come to me spontaneously. I read a book called Prediction Machines about the different ways AIs could be integrated into our lives in the near future. Granted, I did not imagine that this near future was THIS close.



In the world of Artificial Rebellion, these AIs are already integrated into society, so the conflict is not about their acceptance, it comes from the fact that we have a tendency to search for hints of conscience (humanity) in everything, especially in simulated intelligence. I remember one of my critique partners commenting, “Authorbots is such an interesting concept. I wonder what the human writers had to say about it.” We have to wonder no more.


###


What do you think about ChatGPT? Did you enjoy reading my opinion on it? Let me know in the comments!


That's it for today, folks.


See you next post!

Ra.

 

Carla Ra is a scientist by day, sci-fi writer by night. You can check out her anthology ARTIFICIAL REBELLION here.



242 views7 comments

Recent Posts

See All

7 Comments


Marcelo Silva
Marcelo Silva
May 15, 2023

I agree with most of what you have said, but one the things that I feel the urge to address is a current misconception about AI training. Or, in a broader topic, learning. You said "The most obvious example of violation of privacy is the image generators in which the AI has been clearly trained with copyrighted art." but then you also said "(...) did not come to me spontaneously. I read a book called Prediction Machines about the different ways AIs could be integrated into our lives in the near future."

Lets focus on the human component here. You used copyrighted work to train your own brain, ie to learn, inspire and so one. You have never infringed a…


Like
Marcelo Silva
Marcelo Silva
May 16, 2023
Replying to

Indeed, is a gray area and, of course, it is very questionable. If it wasn't, we wouldn't have divergent thoughts on it.


The electric bill mention was to give a perspective of the costs involved... and say that buying a book, or thousands of them, is cheap to OpenAI. They can buy thousands of books, and your argument was that because you bought a book, you could take inspiration on them... that you could learn its contents, be inspired by them, and then create your own. Does that implies that, after buying, OpenAI can use them in their training dataset too? Or, in another way, only buying a book from some author guarantee that you are respecting copyright of said…


Like

chrisrap52
chrisrap52
May 03, 2023

Good objective view. I too see the current AI as a good tool. I have experimented & posted results. Most were impressed with the AI responses. Recently, a novel was written using a few AIs. Good review written by a human, https://www.nytimes.com/2023/04/20/books/ai-novels-stephen-marche.html

Like
Carla Ra
Carla Ra
May 03, 2023
Replying to

Thank you!


Although I actually wrote about novels written by AI, I'm genuinely surprised it became reality so fast.

Like
blue-button.png
bottom of page