Collaborate with AI, Don’t Work Under It — Kabir Narayan Jha
Humans are truly strange. On one hand, we are advancing with incredible technologies like DeepSeek and LLMs, pushing the boundaries of AI. On the other hand, we are gradually losing many of our natural capabilities. Some might disagree with this perspective, but trust me, it is happening.
Do you remember when we used to memorize the phone numbers of 5–10 people? I still recall the days when I practiced writing English grammar and perfecting my spelling. Now, I often rely on AI for assistance. While we are making remarkable technological progress, we are also losing our innate abilities. During my self-reflection, I felt a deep fear of the day when humans will lose many of their core skills. When that day comes, it will bring a crisis.
What’s even more concerning is that we are celebrating this shift. I’m not against AI, nor am I suggesting we shouldn’t use it. However, becoming fully dependent on technology to the extent that its availability dictates everything doesn’t make sense. It will eventually take over our minds and impact our very existence.
As a software engineer, I was a strong supporter of technology in 2012, creating numerous models and witnessing the amazing benefits it brought. However, I also realized that AI comes with hidden problems that are not always easy to recognize or measure. So, I created a personal policy: I will collaborate with AI, not work under it.
For example, many people rely entirely on AI for writing. While I use AI to refine my content, I make sure to write everything myself. When it comes to creative tasks like painting, drafting, and designing covers, I create them on my own first and then collaborate with AI only for tasks that are difficult for me. Even AI has limitations, and for those challenges, I step in to add the human touch.
One example is adding emotions to voice or text — something AI still struggles with in many scenarios. That’s where I find my role, enhancing and personalizing things in ways that others often can’t.
Hope this makes sense.
