Love with Concern: My Passion for AI & ML – Shadowed by ONE Fear

calendar Dec 23, 2025 · 3 min read · AI&ML  ·
Share on: linkedin copy
LLMs with Human in the loop
bg image credit to the owner.

In high school, I fell in love with computer technology through a subject called Computer Studies. I was immediately drawn to two fascinating fields: Artificial Intelligence and Virtual Reality. By the time I was in Form Two, I was absolutely certain; my career path was in Computer Science.

Despite qualifying for and receiving persistent encouragement from friends and family to pursue more traditionally "high-reward" professions like Law, Medicine, or Engineering, I remained committed. I loved Math, Physics, Languages, and Geography. I disliked Biology. But it was Computer Studies that tied it all together for me. It was the one subject that captured my passion for logic, language, science, and nature into one cohesive and fascinating subject.

Pursuing Computer Science at university opened my eyes even further, and quite early, I discovered something even more compelling: Information Security. I became deeply engrossed in the subject, often spending countless hours in the school library reading everything I could find about it. My obsession with hacking(ethical ofc) grew so strong that I nearly neglected other units. The lecturer who taught us Information Security course in second year, Mr. M... became my favorite.

While Information Security took center stage, I’ve always remained a fan of AI and Machine Learning. Like many in tech, I taught myself Python programming, then explored neural networks using TensorFlow (tf), data analysis with Pandas (pd) and Matplotlib (plt), and machine learning concepts using NumPy (np) and Scikit-learn(sk). For my final year project, I built a lightweight ML-based wine recommendation system for e-commerce. It analyzed a user’s past reviews, compared them to similar users, and incorporated comments to generate suggestions, something we now see commonly across online platforms.

A decade ago, in my second year, I would have been thrilled by today’s AI breakthroughs. But oddly enough, I’m not as excited anymore. Perhaps it’s because I stepped away from active research in AI/ML, or maybe it's the overly optimistic, excitement that often feels one-sided, driven largely by non-technical folks who focus only on the "wow" factor and ignore the complexity. I remember hitting a wall while trying to implement Convolutional Neural Networks(CNN) with TensorFlow for natural language processing(NLP). I can’t even recall the exact issue now (I guess it might be context or language evolution) but I bet it’s still a challenge even today – considerable strides have been made of course.

Large Language Models have, without doubt, revolutionized the field. They've been immensely helpful, whether it’s sourcing research papers, discovering security news, trend, projects, resources and stuff, or finding insightful technical blogs. They’ve saved me countless hours I'd otherwise spend buried in search engine results.

My ONLY Fear!

I fear that over-reliance on AI, especially LLMs will slowly erode the culture of deep research, bold innovation, and original invention. We might reach a saturation point, where LLMs stop learning anything new because humans stopped creating anything new. The temptation to let AI do all the thinking for us might kill the curiosity and the human drive to explore, discover, innovate and invent something new!

It’s Not the Ethics, Privacy or Security of AI That Worries Me Most, It’s What We Might Stop Creating. And that’s what concerns me!

References

  1. Nature (2024) – Model collapse https://www.nature.com/articles/s41586-024-07566-y

  2. The Guardian (2025) – Elon Musk on AI data limits https://www.theguardian.com/technology/2025/jan/09/elon-musk-data-ai-training-artificial-intelligence

  3. Wikipedia – Model collapse overview https://en.wikipedia.org/wiki/Model_collapse