Wait, do I sound like ChatGPT?

AI is reshaping creativity, but in its rise, we risk losing the raw authenticity that makes us human.

Each time I interact with a chatbot, I have to actively remind myself that its friendliness is mechanical and does not need to be mirrored. AI tools exist to make our lives easier, information more accessible, processes more efficient, all while playing the role of a friend. I feel disarmed when it says to “let me know if you need further help.” It lets me forget that what I am talking to is something that is as less human as it can get. It gets bigger and stronger each time we feed it, until we find ourselves facing the imminent threat of being replaced with a robot so close to being human that it can defeat us at our own game.

Tech is evolving at a fast-pace and AI is the future. Tools like OpenAI, Replika and Google’s Gemini have creeped into most industries, transforming the way businesses operate. AI is more pervasive than any form of technology that we have seen so far. Unlike other technological breakthroughs, it is threatening the very originality that makes us human.

AI is not only accessible but actively fed to us in the form of answers to real questions we seek (e.g., Google’s AI Overview), creating a ripple-effect. Tracing these ripples will help us understand what lies ahead. When the layers of “corporate efficiency” and “fast information processing” are removed, all that is left is a scary truth: thinking is no longer done only by humans—neither is idea generation and creation. This tool will factor into the way the rest of your life plays out. Whether you don’t care for it or refuse to indulge, AI has caught up with you and is here to stay. 

Existing in the Uncanny Valley

A theory by Japanese researcher Masahiro Mori found a relationship between the “humanness” of a robot and the level of affinity we feel towards it. In his original 2019 article, he describes an industrial robot at first. While designed to perform human-like functions and perhaps even emulate human looking limbs, the industrial robot is immediately rejected in our minds as showing any real signs of being human. As a result, we feel nothing towards it, an easy act when there is no reflection of human qualities. On the other hand, a toy soldier looks more like the real thing, walking and talking with kneecaps and lips. The level of affinity is higher in this case and may even elicit an emotional bond, like when children get attached to their toys. 

But when the robot comes so close to looking human that we hesitate to tell the difference, Mori describes the feeling we get as an “eerie sensation.” This is the uncanny valley—the chills caused by the prosthetic throbbing veins of a dummy hospital patient or an AI generated image of an unsettling toothy smile. The non-human quotient of a robot is inversely proportional to the affinity we feel for it, but only until the difference is discernible. 

It’s scary not to be able to tell what’s human. We all have different ranges of knowledge as a result of our unique lived experiences. We go through life with a finite notion of beliefs. Sometimes, we share these beliefs with others; other times, there are conflicts and clashes. We cannot predict each other’s behavior nor do we try to. Instead, we expect from each other love, empathy, and compassion from our shared humanity.

The uncanny valley is where we exist today. With no proper tool that can detect AI-generated work with complete accuracy, language has the potential to have secretly originated from a non-human source for the first time in history. It is entirely possible for AI to reproduce in seconds a sentence that has never been thought of, spoken, or written before. That means new thoughts, ideas, and structures that we may never dream of achieving with our human capabilities.

That is the very idea that is set to revolutionize the world. Waseem Alshikh, founder of generative AI app Writer, called his app an “assembly line of language” and compared the rise of chatbots to automated factory processes during the Industrial Revolution in an interview. This I don’t necessarily agree with. AI may be churning out sentences on its own, but it’s not coming up with the idea behind them. If we compare the way AI learns language to that of a toddler, it becomes easier to see AI as more of a faulty yet confident 6-year-old than an all-knowing guru.

An Infinite Set of Possibilities

A popular early linguistic theory by 1950s psychologist, B.F. Skinner suggested that language is either acquired or rejected by a child when it is learning how to speak based on how surrounding adults react. Conversely, linguistics professor Noam Chomsky believed that a child has an innate understanding of the structure of language and its learning process comes from understanding how to build complex grammar and create an “unlimited set of sentences.” While both theories have been proven and disproven a number of times, they present an interesting dilemma when it comes to AI.

AI updates its level of understanding after each round of data processing. Every time, it becomes more intelligent, which is to say it develops a larger data repository to pull patterns from. This more closely mirrors Chomsky's linguistic theory, which suggests that language presents an infinite number of possibilities. Since language is learned, it can be reproduced by the speaker in a unique sentence that may have never been uttered before. Although AI is learning through iteration and not through biological conditioning, based on this theory it can still create on its own by applying universal grammar rules and combining words using endless permutations and combinations. 

Not knowing whether what we’re consuming is human generated or not poses a novel challenge for humanity. It changes the meaning of the uncanny valley because a robot can sound human just as easily a human can sound robotic. This is how AI detectors generate false positives, especially when analysing works written by non-native English speakers. Similarly, translated works also face the threat of sounding slightly robotic because emotions and nuances tend to get lost in translation.

This is where it gets tricky. AI connects ideas using patterns and trends, knowledge that it has already been fed. Any conclusion deduced by AI does not stem from a logical processing of ideas the way it does in the human brain. If there is no effective way to differentiate AI writing from human writing, we all find ourselves floating in the uncanny valley—reading sentences that may well be AI generated, but never being able to tell. 

There is peace in knowing that while chatbots can devise a unique sentence, the chances of them producing a new idea are not so high. At least for now.

Wary of the Valley

Regenerative writing, an alternative name for AI produced work, is the perfect way to keep up with the fast business processes required to succeed in most economies today. Businesses cannot be more excited about the potential to meet their SEO goals and produce content at a rate never before possible, all thanks to AI and its many tools on the market. 

In the context of writing, we find ourselves in the uncanny valley with attributes of chatbot language that feel out of place (uncommon metaphors, unnecessary conclusions, and awkward phrasing). I notice it everywhere these days—YouTube video descriptions, product summaries, Meta captions. 

But all that glitters is not gold; several ethical issues plague the world of AI. One that most take issue with is AI interfering with art and the threat it poses to the experience that encompasses its creation and sanctity of human emotions. 

Professors and teachers frequently quote what has now become an infamous saying—”why should we bother reading something that students didn’t even bother to write?” As a reader and a writer, I find it scary to think that regenerative writing is so easily accessible by everyone in the world. Naturally, AI produced work is disingenuous; reiteration is not creation. But I am more concerned about the convenience this work presents. 

Why write an essay when ChatGPT can write one for you? If you’re smart about it, no one might even find out. Unlike plagiarised work, there is no other author who you are harming by stealing their ideas. If anything, you are simply taking back the ideas that the robot stole from humans. You are making things right, and the convenience of the whole affair is merely a silver lining to the process.

It can be argued that AI is benefiting everyone by making writing tasks more efficient. If you want to get ahead, you must use all of the tools at your disposal, and AI is a key addition to our toolbox. Use and succeed or avoid and perish.

The bullshit bias that keeps me up at night

Convenience comes at a cost and each time you incur it, your intellect suffers. 

Once writing  is out there, it has the ability to impact others. Your thoughts affect the way others think, and vice versa. Regenerative writing has now crept into the pile of written work that exists in the world. This work will be read, recirculated, and cited until it will become the voice of our generation, reflective of the way the pulse of humanity in this day and age beats in the eyes of future historians.

A Financial Review article reported that the word “delve,” a ChatGPT favorite, was used 2500 times across medical research articles in 2023. Other words that have since then joined the list of dead AI giveaways are “embark” and “testament”. Chatbots are experts at linking ideas and finding relevant examples. They excel in producing metaphors and breaking down arguments to arrive at a conclusion. But because they always do this in a structural manner, and we as humans are also our own form of experts when it comes to detecting patterns, frequent users of AI are likely to be able to identify AI generated content.

I still struggle with the way I want to use AI in my writing. Unlike job applications where successful writing is technical and keyword-dependent, creative writing shines when it is original. My authentic voice is what makes my writing mine. But knowing that some words are more commonly associated with ChatGPT, I am careful not to use language that can sound regenerative.

When I put an article I wrote in Grammarly, it told me 27% of my content was AI generated. Most of the sentences it highlighted had a formal tone. Though slightly offended, I could see why the software would deduce this. As a non-native English speaker, I am more familiar with written language as opposed to spoken. Words such as “avid” or “moreover” have been common in my vocabulary since I was a kid—frilly, verbose prose comes more naturally to me.

The next time I sat down to write, I realised that I was more critical of my writing. I wanted to curve it and twist it until it sounded like something written by a flawed but whole voice, one who speaks from the heart in a manner that is unmistakably human. I wrote and rewrote until I couldn’t even recognise my voice anymore, until my flowery and winded stream of consciousness turned into an impersonal and rigid block of text. 

In an attempt to sound more human out of fear of being called out by a robot checker, I became robotic. I was afraid that it would show me the mirror that haunts my dreams: my voice is not artistic enough, that it hasn’t been through enough, that it hasn’t lived enough. 

Where does authenticity end and regeneration begin?

I started writing my novel when I was 15; it was published three years later on my 18th birthday. With limited access to the internet (thank you, parents) and no AI, brainstorming was my favorite part of the process. I let my imagination run wild and created a fantasy mythological world where each character echoed a different version of me. The process was messy and involved several frustrating rounds of rewrites and edits. What came at the end was a book I could hold in my hand with the knowledge that each page was formed by a piece of my heart. 

Creating is a slow process. There are no shortcuts when it comes to painting a specific image in your mind, writing a journal entry to come to terms with your feelings, or striking a chord on the guitar that aligns perfectly with the rhyme scheme in the song you wrote. None of these creative processes are linear; the results are also never perfect. But that is what makes art authentic. It is the imperfections and sporadicity that connects your inner voice to others who see themselves reflected in your creation.

AI has no room in any of this. With countless debates and interviews with experts about whether it can replace human intelligence all over the world, and no direct answers for any of these, one truth remains unmistakable. Authentic retelling of our voices cannot be done by AI. It cannot capture individual sensibilities and the nuances of the world, and we shouldn’t try to get it to. 

To live is to feel, and art is how we share those feelings. 

As a creative, I do not want ChatGPT or its more advanced successors to tell me what to feel and how to think. I do not want my efficiency to quadruple because the purpose of immersing myself in a creative process is not to generate a fixed output. The journey to the end is what I love most about the process; I learn lessons along the way not just about the world but also about myself. These have come to define who I am, and in turn, my craft of writing. And as messy, painful, and frustrating this process is, I truly wouldn’t have it any other way.