When AI becomes your voice: A personal reflection on over-reliance

In a world where tools are becoming increasingly smart and accessible, where is the line between enhancement and dependence?

Over the past year, Gen AI has become my go-to assistant; coming up with content, debugging code, exploring new ideas, or simply looking for a second opinion on how to rephrase a sentence for an email. As part of my daily workflow, GPT has boosted my productivity and helped me overcome the writer’s block (being unable to think of what to write or how to proceed with writing) when documenting manuals and code.

But recently, I faced a strange problem, one that made me stop and reflect…🤔

I was applying for a job and, as part of the instructions, applicants were told not to use AI to generate content in their applications. I took that seriously. I sat down, brainstormed, outlined my points, and wrote my cover letter from scratch. No prompts. No templates. Just me.

Out of curiosity (and a bit of paranoia), I ran my final draft through a few AI detection tools.

Result?Aaaah!

Over 90% AI generated. It was then it hit me: I’ve absorbed the language and tone of AI.

“Kimeniramba” (a popular Kenyan phrase meaning being caught in a difficult situation due to the consequences of your actions) was smiling at my face. I am reaping what I sowed. Effectively. My one Nigerian friend would tell me that my “village people have succeeded”, but, well atleast my village has nothing to do with AI 😅

My writing had started to mimic the patterns, structure, and tone of AI. Clean, formal sentences. Balanced arguments. Predictable phrasing. It wasn’t plagiarized. It wasn’t even assisted. But it no longer sounded like me.

The invisible influence

Generative AI doesn’t just generate content, it shapes how we think and speak. Especially for those of us who write on a regular basis, there’s a shift that happens when you’re constantly exposed to how AI “talks”.

You begin using its transitions, styling (emojis like this popular one🚀), you adopt it’s rhythm (short paragraphs, clear bullet points), you unconsciously censor your tone and raw voice in favor of polished structure.

And so, the birth of dilemma and dilemmas:

1. Do we change how computers think and talk, or do we change how we think to sound less like them? Have we absorbed AI or has AI absorbed our own language? Thinking about it… we are generating new content from AI that has been trained using our old content.
2. How do we maintain our individual voice and thoughts while using AI to enhance productivity and creativity?
3. Will institutions begin to penalize content, or employers reject our applications, mistaking it for AI generated? (…side eye-ing my lecturers 👀)
4. What happens when our own thoughts are no longer distinguishable from machine outputs?
5. Do we need better AI detectors or do we need to talk about AI literacy, digital habits, and how we train ourselves to set boundaries?
6. Do we need some additional/enhanced(if in existence) “creativity-first” guardrails that will prompt the person for their own creative input before jumping right into the AI output impulsively? Something like GPT asking you to write the content yourself before it gives you its thoughts and additions on the topic.
7. Or are we also supposed to rely on AI humanizer tools?

… or we should go native and start publishing content in our native languages.

This experience left me with more questions than answers. I’d ask myself again: In a world where tools are becoming increasingly smart and accessible, where’s the line between enhancement and dependence?

Finding the balance

Personally, I’m not giving up on AI, I’m infact, doing more research on this subject and hoping to engage in global talks, discussions and innovations in AI, to address its impact in the society: social, ethical and legal wise.

AI is still and will remain one of the most powerful tools in my toolkit, like Thor’s Hammerrrr. But I’m also learning to draw boundaries and learn:

–       To write without help first

–       To read more human-written work; novels, essays, conversations…so as to rediscover what natural and not-Gen AI thoughts sound like.

As I am pursuing my masters in AI, I am getting involved with organizations such as Internet Society (ISOC) and Kenya ICT Action Network (KICTANet) to engage and learn more on policy advocacy, ICT governance, digital rights and current issues on the internet and AI.

My hope is that we find a middle ground where AI is a collaborator, and we remain the authors of our own voice and creators of our own work. Because soon, we won’t be able to use our brains creatively, or perharps even use them at all (when we bring automation in the picture). We will always argue about the need for efficiency and speed, well, in exchange for our natural thoughts. And thats how the coming generation is at a greater risk.

… If you’re wondering whether you sound human. Share your thoughts and let’s see if “Kitakuramba”😂😂(just kidding)