Why does a computer not understand that rice is healthier than Maggi

iSRL is an open research lab building IFID — the coordination layer that makes Indian packaged food ingredient data legible to machines, regulators, researchers, and industry.

December 2024 during a hackathon, the computer declared very much proudly ‘Maggi is healthier than rice’

But oh oh, we know it is about to get slapped by every grandma in 5km radius.

It started with a simple question: why does a computer not understand that rice is healthier than Maggi?

Not a trick question. A real one. The answer — once you start pulling on it — is that the systems downstream food data runs on have no shared way of understanding what an ingredient actually is. The same thing appears under dozens of names across labels, regulations, trade records. “Groundnut oil,” “refined groundnut oil,” “arachis oil” — the same source, no connection between them that a system can see.

Multiply that across every ingredient in every category of Indian packaged food and you have a landscape where data cleaning is half the job, compliance is guesswork, nutrition studies are built on shaky ground, and no one can ask a clean question and get a clean answer.

“Hey by refined wheat flour here, just to be sure you meant maida, not the atta chappati one right?”

This is not a fragmentation problem. The diversity is real and it deserves to exist.

A farmer in a village should not have to adapt to a standard that was built without him.

A brand using tamil language to list the ingredients in the food packet is making it readable to the very consumers they work day and night to make happy.

The regulator who doesn’t force everyone to use same language isn’t them being inconsiderate, it’s them making space for the inherent diversity and richness of India.

So the key is, everyone is trying their best to make sure the food we eat is healthy, tasty and makes us feel joyful better than otherwise.

The one writing this behalf of everyone involved with this, I have no nutrtion degree. I am not a nutrtionist1, we are not here to judge, but we are trying to make the data judge-able1 — to regulators, researchers, brands, and anyone else who needs to ask a real question and trust the answer.

1 Please don’t ask my biology teacher in high school of how well I learned biology. But data is something I am good at…let’s just say so please?

What’s missing is not uniformity — it’s a coordination layer. A way for backend systems to talk to each other about food without flattening what makes each ingredient, each tradition, each regional name meaningful.

That’s what IFID2 is building. A deterministic ingredient substrate3 — a canonical reference that makes the ecosystem legible to machines without requiring the ecosystem to change.

2 Indian Food ID or if you want us to sound more ‘precise’, it’s Indian Food Informatics Data. It’s the term for what we are doing apparently. Yes we googled it.

3 Think of this as UPI for food?

We’ve been testing its implications as we go: trade blockage simulation, compliance and R&D acceleration for industry, nutrition studies where the data actually holds.

The outputs are neutral and open under CC BY 4.04

So yeah, you’ve found an open research lab building the data infrastructure that lets experts ask questions Indian food systems couldn’t answer before.

It’s an open coordination point. The work draws on regulatory expertise, food science, data standards, legal analysis, nutrition — no single discipline has the full picture and none is more important than the others. Every person who’s come into this room has brought a worldview that changed something we thought we understood.

If this problem pulls you in, there’s a place for you here.

We couldn’t have coined the term Regulatory Delta Analysis if it wasn’t Lalitha (aka me :>) trying to find a way to explain to Sai how we can maybe analyse two sets of law to find the underlying directions, motivations and constraints to use it as data, which became

We couldn’t have known the parallels and how similar systems in completely different industries do this without Thej saying so here.

We couldn’t have thought of how certain aspects may not be realistic for a small F&B businesses if not for Dr.Swati pointing it out here.

We were working on this problem from the food side and found out Google Research had converged on similar structural thinking from a completely different direction — we couldn’t have done that without all of our worldviews and unique intelligence that we all bring in.

The point is, everyone who contributes bring in the inherent intelligence you have that we otherwise can’t, so if you have something to contribute to serve this problem with us, you belong with us.

You might want to be in the trenches — contributing specialised expertise as a subject matter expert directly to the research, getting credited and published for what you build

You might want to sit at a remove and break things — reviewing outputs, finding where the reasoning fails, pushing back as an advisor.

Or you might be someone who connects — who sees across domains and wants to be the coordination point itself, as a systems researcher who holds the different threads together.

The problem is the invitation. If it’s pulling you, come in.

Here’s how.