How Robots and AI Are Changing Farming

Farming Techniques
10 followers
4 Video Views·Nov 5, 2025
Robotics and AI are reshaping how food is grown. From advanced robotics and drone technologies that promote more efficient watering and weeding systems to complex seed and soil analysis that makes the food chain more efficient and resilient, AI is influencing food systems. We speak with an innovative robotics farm equipment company and visit a research lab to see how AI is factored into the future of farming.
00:00 - AI on the Farm
01:28 - Agriculture AI in Context
02:05 - UC Davis
04:01 - Digital Green
05:18 - Environmental challenges of AI
06:43 - Verdant HQ
Daniel Alameda is a third generation farmer in the Salinas Valley,
where he grows thousands of acres worth of broccoli, cabbage and lettuce, among other things.
Like with any farmer, the scourge of his operation is weeds.
And about a year ago, Alameda turned to a cutting-edge solution.
An AI-powered weeder.
The device, which attaches to the back of his existing tractors,
uses high-resolution cameras and sensors to automatically detect unwelcome plants
and spray them with a precise dot of herbicide.
So what the driver sees in the cab is real-time footage of what's happening.
The crop that's there and the weeds that we are trying to remove.
The driver sees the machine analyzing the data and then taking the shot.
You're seeing decisions made in real time.
We definitely had no AI not that long ago.
The owner-operator-manager was the AI.
We were making the decisions.
But we realized quickly about 10 years ago that we could use software to make better decisions.
We are trusting AI to make us better, more efficient and faster.
We're starting to kind of notice AI getting a lot more penetration on farms across the country.
Farmers are using things such as drones, autonomous plows,
crop sprayers that can tell the difference.
Using AI between a weed and a crop.
And a lot of them are using data collection to just suck up all of the information from the farm.
John Deere is the biggest farm machinery maker and they have a goal by 2030
to be able to farm 100% autonomously.
So I think this stuff is here to stay.
And there's really not a lot of other ways to just kind of get it all done.
At UC Davis, researchers are using AI in another way.
To better understand individual crop genes.
So they can make them healthier or more resilient to changes in the environment.
What you're seeing here, the first prototype in half a day was this thing.
A bird box right now.
It's one thing to collect the data.
That part's actually kind of easy.
You go out in the field, you collect the data.
But then you generate this massive data set.
It's like imagine if you took your camera instead of just taking red, green and blue,
which is what it does now, took another 400 bands of light.
It's this cube of pictures.
And so what you end up with is this massive data set that has to be processed somehow.
That's where artificial intelligence comes in.
Scientists at Davis are training computer models to detect everything
from the size and shape of a flower to the direction that a leaf is pointing.
Plant breeding is a really interesting application for AI.
Let's say we're working on a bean.
In this example here, what we have is over 330 different genotypes of beans
that have been planted out in a field.
And so what you're looking for is like maybe some of these have really big beans
or certain patterns or the protein content is high or the fat content is high.
It's just really hard to measure and see and score by humans on the ground.
And then if we can process all that data using machine learning and AI,
it allows us to extract those complex traits.
I think farmers from 50 years ago,
if they saw what AI could potentially provide them today, I think they'd be very excited.
Using computer vision, we can rapidly accelerate that process of searching
through all these possible children, right?
The traits that we care about.
And so it can take us from taking 30 years to breed the crops that we want
to maybe taking three years.
The fact that we can then take something that seems so state-of-the-art and advanced
like AI machine learning and bring those two together,
I think there's something really magical about that idea.
What's the right time of year for me to begin setting up my raised bed nursery?
The best time to start setting up your raised bed nursery for
transplanting paddy saplings...
The AI tools being created in collaboration with the team at Davis
aren't just about using AI to process and analyze data.
Rik and Gandhi has been building a large language model app
for farmers in developing nations.
So the small scale of farm we work with are primarily in South Asia
and in Sub-Saharan Africa, specifically in India, Kenya, Ethiopia, Nigeria.
You could say that what we're building is a type of chat GPT for farmers,
but they're not looking for generic Wikipedia or chat GPT answers.
They're looking for information that is proximate to them and location specific.
There's a bunch of challenges that we face.
There's a need to support all the various languages and dialects that farmers speak.
They don't use scientific terminologies like kilograms or percentages or chemical names.
We need to be able to train these systems for speech-to-text recognition,
for their local colloquialisms and vernacular,
and also not just speech-to-text, but also text-to-speech,
so that farmers who often have low literacy can be able to interact with these systems.
Of course, AI is not without its own environmental costs,
something those working in farming are all too aware of.
Training a large-language model can use up millions of gallons of water,
and data centers dedicated to artificial intelligence
can do real damage to the land and power grids around them.
These farming communities have been so often left out of our
collective improvements of digital technologies,
and so I think that we also have to ensure that there is inclusivity
to be able to access this technology,
even if there might be an expense of energy and cost for them to do so.
It's about providing them more of a level playing field.
How AI impacts the environment and is that relevant to agriculture
is a really important one we should be asking.
On the research side, we're using, honestly, very small amounts of resources,
like a few GPUs, eight, to train these,
whereas some of these large-language models are using more like tens to hundreds of thousands
for many months, as opposed to eight hours.
So I don't worry too much about our personal impacts, I know what they are,
but I think like when you think of scale, like of the industry,
I definitely worry about the energy effects of AI.
We are at the beginning, it's moving quickly,
and I think we come back and revisit this three to five years from now,
we're going to be pretty blown away with the impacts that it's had already.
Okay, so what we're going to show here is,
so this is a live feed from the robot, right?
We have dye here to see the shots well.
The weed-killing robot that Daniel Alameda used on his farm
is produced by a company called Verdant Robotics.
Gabe Sibley, who spent his early years working on the Mars rover,
founded the company in 2018.
The machine knows where it is in space,
so it's got spatial AI that allows it to solve the where problem,
where am I?
It also solves the what problem, what am I looking at?
Is it a carrot? Is it a weed? What type of weed?
Initially, people thought that autonomous navigation
would be a really valuable thing to bring to market.
The thing that you have to understand is that
it's not the navigation of the tractor that's the job they're worried about.
It's the hundred people walking behind the tractor
that are opting to go work at Starbucks or Walmart
or in a data center labeling images.
That shift in labor is the one that's causing them heartache.
Efficiencies we need in ag are substantial.
In the next 50 years, there are folks that say we need
to grow as much food as we've grown in the last 10,000 years.
Half of the habitable land on Earth is already under cultivation.
We have to do it more sustainably and more efficiently.
What you'll hear over and over is just the tight labor market.
There's never enough sort of field workers or machinery operators.
A lot of farmers would not recommend their children
staying on the farm and would kind of push their kids to find careers elsewhere.
But these machines are also an opportunity to attract,
you know, the younger generation that's interested in video games
or that's interested in new technology.
And that might be a way to sort of start clawing back
some of the folks from the city back to the farm.
We have just stuck our big toe in the water for the first time.
And it seems like it might be inviting.
It's also a little scary.
We don't know exactly what's going to happen.
As a species, as we start to automate all types of work,
really not just agriculture,
you've now entrusted all that work to a machine.
And so we better make sure that we are still in control.
The difficulty then is like if the switch goes off,
do we remember how to farm?
We better.
