This all began one evening when I asked my husband:
“What is AI?”
He, being the natural technologist in the house, gave me an explanation that sounded like the beginning of a sci-fi movie — CNN, RNN, deep learning, Transformers, machine reasoning. He spoke confidently, enthusiastically… and I just stared at him thinking, excuse me, “what language is this?”
He gave me the gist of it, but the jargon did not land. It didn’t even walk past me; it flew over my head. But it sparked something. I was curious. I did what I always do when I don’t understand something — I kept asking questions. I know that to make sense of it, I needed to relate it to something I knew. So, I asked him, “Can I say AI is like a brain?”
He smiled and said, “Kind of — but only on the outside. It has ‘neurons,’ but they’re just math functions, not real cells. It recognises patterns, connects information, and gets better the more data it sees. That’s why it can predict things or even generate text and images. But it doesn’t understand, feel, or think. It just follows patterns — really, really fast.”
And that’s when it clicked to me – the cue for me to relate AI to “the way humans learn language?” That was my anchor — my English degree, my experience teaching early literacy to trainee teachers, and my personal journey teaching my daughter to speak, read, write, imagine, and express herself.
- CNNs learn to recognise shapes and patterns → like children learning letters
- RNNs learn sequences → like children learning phonics, grammar and sentence flow
- Transformers learn context and meaning → like children understanding stories – begin self-correction and making sense of context and meaning.
- Generative AI produces creative outputs → like children creating stories, drawings, ideas.
The analogy may seem a bit far-fetch at first, but it makes sense to me and the moment he explained AI using the language of human learning, everything clicked. Suddenly AI wasn’t this mysterious, robotic monster. It was something familiar. Something I could relate to.
It also reminded me of something important: humans learn best when learning is made meaningful. When someone meets us where we are. That is something technology can never replace — the human ability to turn confusion into understanding by connecting something new to something known.
And so, the more I learned about AI, the more I saw the parallels with child development — especially through the milestones I watched my daughter achieve.
The Early Stage: Pattern Recognition
When my daughter was very young, I was amazed at how she absorbed the world around her. She was surrounded by books, alphabet toys, labels, and other print-rich environments. Reading and exploring print was our shared activity, alongside playing with toys. I still remember when she was about 22 months old — we were on the bed, flipping through her favourite book on animals, when she suddenly read the word “Lion” as we pointed to it. Encouraged, we moved on to another word — “Roars” — and she read that too. “Roars…” she said. We were both stunned! She was not yet two, and she could read. By the age of two, she was already recognising letters and reading more books with simple words, scanning letters, connecting shapes to sounds, and recognising familiar sequences.
I relate this stage of learning “Print Awareness” to Convolutional Neural Networks (CNNs) in AI. CNNs are designed to detect patterns in images or text. They don’t understand meaning yet, but they scan, observe, and link visual cues together. Just like at “Print Awareness” stage, a child notices:
- this shape is “A”
- this curve is “B”
- print is read from left to right
- this pattern appears in words
- this colour means something familiar
In schools, CNNs can scan handwritten assignments, diagrams, or worksheets, quickly identifying key features for review. Both child and AI start by absorbing and recognising — understanding and meaning come later. Exposure, repetition, and guidance are key in both cases.
Sequence and Context: Learning Rules
As children grow, their learning becomes more structured. Just as my daughter could recognise letters and read simple words, she was also tuning into sounds. She learned that “c-a-t” makes /cat/, and that some letters sound different in different words. This early phonological awareness helped her predict and decode words.
By age three, she had begun forming her own sentences. She experimented with grammar, understood simple sequences, and anticipated what might come next in a story. I still remember her coming home from school and telling me, “Ibu, I ate candy at school today. I want to eat ice cream at home.” The fact that she intuitively used ate — the past tense — to describe something that happened earlier shows how she was internalising rules of sequence and time.
I see this similar to how an RNN processes sequences to anticipate the next item in a pattern. They process information in sequence, remembering what comes before to guess what comes next. In language, this is what happened at “Phonological Awareness”/ “Decoding” stage – a child began to understand rhyme, syllables and sound patterns, meaning:
- What word fits this sentence?
- What sound completes this pattern?
- What comes after “once upon a time”?
In education, RNNs help with context-aware feedback — tracking progress, guiding writing, and supporting problem-solving. Like children, AI improves with practice, feedback, and repeated exposure. Both need structured guidance to connect patterns with meaning.
Understanding Meaning: Transformers and Deep Learning
By four, my daughter wasn’t just reading words — she was reading stories. She could finish whole books, imagine different endings, interpret characters’ emotions, and even create stories from pictures. She was expressing herself through play — making cakes or figures with playdoh, turning imagination into tangible creations, connecting ideas, interpreting narratives, and experimenting with expression.
This stage aligns with Transformers and deep learning models in AI, which can analyse long sequences of information, attend to multiple inputs simultaneously, and generate coherent and meaningful outputs. In the classroom, this allows AI to summarise complex texts, connect ideas across subjects, and provide explanations that make abstract concepts accessible. Like a child weaving meaning a through life experience. AI links meaning statistically through patterns in massive data (without human intuition, empathy, or curiosity) but the outcome looks similar.
Creation: Generative AI as the Adolescent Stage
Once a child internalises patterns, sequences, and context, the ability to create independently begins to emerge. My daughter’s storytelling, imaginative play, and hands-on crafting reflect this perfectly. She can invent stories, transform prompts into drawings, and craft playdoh figures, bringing her imagination to life. She is only four, and yet I can already see the beginnings of her creative identity forming — her ability to turn raw input into something uniquely hers.
Generative AI mirrors this creative stage. Tools like:
- ChatGPT, Grok, and Gemini generate text
- DALL·E, MidJourney, and Stable Diffusion create images from textual prompts
- Runway or Luma AI produce videos or animations.
These tools produce outputs that appear imaginative, even artistic. Yet, unlike a child, AI’s “creativity” is not curiosity-driven — it is pattern-driven. Still, it opens doors for humans to express ideas faster, more freely, and more boldly.
The way I see it, in education, Generative AI allows students to experiment, ideate, and create, much like a child turning ideas into tangible expressions. It becomes a co-learner, amplifying creativity, providing inspiration, and supporting exploration, while human guidance ensures that creativity remains meaningful, purposeful, and ethical.
Autonomy: Agentic AI and Independent Learning
The final stage is autonomy. Just as a young adult can act independently, make decisions, and plan towards goals, agentic AI can also make decisions and take actions towards specific objectives.
My daughter is only four, so she is just beginning to experiment with decision-making.
Just today, I asked her what she wanted to watch on TV, and without hesitation she said, “I want to watch Mr. Bean Holiday.” When I asked, “Why?”, her answer didn’t just explain her reasoning — it revealed her feelings: “Because you are busy. I want to watch something fun, and Mr. Bean always has silly face.”
These small moments — choosing which story to read, deciding how to build her playdoh figures, or imagining her own endings to a familiar story — show early glimpses of independence. And none of these moments happen in isolation. Her autonomy is shaped by her exposure, our guidance, and the richness of her environment. Her choices emerge from her experiences.
Autonomy in AI works the same way. It is not independence without oversight. Like a young adult guided by mentorship, values, and experience, agentic AI requires human supervision to ensure it acts in alignment with fairness, ethics, and purpose.
In education, this means AI can help students explore independently, suggest learning pathways, or make small decisions within simulations or projects. Yet it is never a substitute for a teacher, mentor, or human values. Autonomy in AI is most powerful when it is guided — because no matter how fast or “smart” it appears, its judgement is never lived, felt, or human. Ultimately, it is still human experience, ethics, and wisdom that shape the way AI acts.
Truth is, AI and the brain is not an apple-to-apple comparison. AI grows exponentially faster than any child. What takes a child months or years to master, for example, reading fluently, writing stories, or solving complex problems — AI can process in seconds or hours. CNNs analyse millions of images in moments. RNNs and Transformers process long sequences of language almost instantly. Generative AI can produce stories, images, or videos in the blink of an eye.
AI is, in this sense, a child prodigy on steroids. Speed and scale are astonishing, but unlike a child, AI lacks intuition, curiosity, emotional understanding, and lived experience. These human qualities remain essential, which is why education must remain human-centered.
What does it mean by this? “Humanising Learning Through AI”?
Raising my daughter has shown me that learning is never just about exposure to knowledge. It is about curiosity, creativity, resilience, and ethical reasoning. Means, as parents we can create the richest environment, provide every resource, and guide every step — but if she herself doesn’t have the curiosity to explore, the courage to try new things, or the desire to understand, all effort becomes meaningless.
Learning requires will, and that will is uniquely human. AI does not have it. That is what sets humans apart.
Much like, in education, AI can amplify human potential, but it cannot replace the teacher — the mentor who inspires curiosity, nurtures critical thinking, and models empathy. AI can support, guide, and extend learning, but it cannot feel, reflect, or care. Its strength lies in partnership with humans, not in autonomous learning.
AI can help:
- Personalise learning: Identify gaps, recommend resources, and guide students along tailored pathways.
- Amplify creativity: Encourage storytelling, problem-solving, and imaginative play — much like a child turning prompts into stories or playdoh creations.
- Reduce administrative load: Automate grading, track progress, and provide instant feedback.
- Promote critical thinking: Encourage students to evaluate outputs, reason independently, and make informed decisions.
Used thoughtfully,
AI is a partner in learning.
It can articulate ideas and support creative processes, but it cannot act autonomously in the real world or replicate human emotional intelligence. Leveraging on it means we are enabling students and teachers to focus on what matters most – curiosity, creativity, connection, and meaning. Its value is unlocked only when paired with human guidance, reflection, and oversight.
I hope for a future where AI amplifies human potential without replacing the human touch, where students can dream, explore, and create, guided by both technology and the wisdom, empathy, and values of human educators. AI can be fast, powerful, and generative — but it is the human heart, guidance, and insight that give learning its meaning. That is the future I aspire to — for education, for AI, and for the next generation of learners.
Leave a Reply