Can AI Make Biomimetics Smarter?

I've spent the last few years building materials the standard way of mixing, testing, tweaking, and repeating until something finally worked. That's how most materials science still gets done: trial and error, one sample at a time, guided by intuition built up in the lab.

Lately, though, I've started looking at this process from a completely different angle: what happens when AI and machine learning enter the picture?

Biomimetics has always asked what we can learn by copying nature's designs, bone that self-heals, shells grown from seawater with no furnace, silk stronger than steel by weight. The catch is that nature doesn't prescribe the design; scientists have had to reverse-engineer the "why" behind every design, the hard way.

That's exactly the part AI is starting to speed up. Instead of testing one polymer or peptide at a time, machine learning models can scan thousands of candidates, learn the hidden patterns linking structure to function, and predict what might work before a single atom is ever synthesized.

It's changing how I think about experimentation itself, less about guessing and testing one thing at a time, and more about learning nature's playbook faster than ever before. I dabbled with AI & ML while creating the material for filtering out oil & microplastics, with inspiration from the lotus leaf. More to come on AI & how it helps in material science.