[ Question ]

Could next-gen combines use a version of facial recognition to reduce excessive chemical usage?

by ddavies1 min read13th Jul 20204 comments


Personal Blog

I came across an assertion that next-generation combines are / will be able to load up with herbicides, pesticides, fungicides, and fertilizer all at once, then go through a field, and the cameras on the combine would not only identify whether a plant is a weed or a crop that's supposed to be there, but also whether to apply herbicide/pesticide/fungicide/fertilizer. This could potentially greatly reduce excessive chemical usage, thus by extension reducing run-off, algae blooms, etc. and the effects thereof. It could probably also determine whether the plant needed more or less water, thus potentially optimizing water usage. There would, of course, be risks inherent in and downsides to this technology - it very much fits into the Borlaug school of thought (see The Wizard and the Prophet).

Does anyone have more info on whether this tech is feasible and, if so, whether it's being developed?

New Answer
Ask Related Question
New Comment
4 comments, sorted by Highlighting new comments since Today at 10:34 AM

Would you mind expanding on why you think this is related to doing the most good? 

Also, what inherent risks are you referring to? (do you say that there are risks in any innovation?)

Sure, sorry for not having spelled it out in the initial post. It's related to doing the most good in that overuse of pesticides, herbicides, fungicides, and fertilizer have serious impacts on the health of soil, water, people, insects, birds, amphibians, and mammals. Further, by reducing soil degradation and increasing the efficiency of farming, one could also reduce the pressure to clear forests for new farming zones while simultaneously increasing the earth's carrying capacity. Additionally, reducing fertilizer overuse could mean cutting back on the production of fertilizer, which is energy intensive and thus contributes to climate change.

Borlaug's Green Revolution is lauded as having saved millions of lives ("Norman Borlaug conducted research into disease-resistant wheat, helping to bring about the ‘Green Revolution’; he has been credited with saving hundreds of millions of lives" Introduction to Effective Altruism); this technology, if feasible, could be a big step in addressing many of the negative side effects of the Green Revolution, including overuse of fertilizer, water-clogged soil, damaged ecosystems, and polluted waterways. This therefore seems like an exciting possibility for people who care about the long-term future.

At the same time, it holds some of the risks that the Green Revolution held: such technology is likely to favor large, industrial farms. From what I understand, the Green Revolution led to further concentration of land in fewer hands, having negative effects on smallholder farmers who either found themselves unable to compete with farmers using industrial tech or were dispossessed as their farmland now became more valuable. Further, greater automation of agriculture could increase a country's fragility if the systems were to be hacked and sabotaged. Famine is already used as a weapon of war, so this doesn't seem so outlandish - though admittedly I think a lot of US agriculture is already controlled remotely. Further automation of agriculture could reduce the points of failure compared to an offline system.

Finally, I posted my question here because the EA community seemed a good bet for finding an intersection between people who care about environmental systems, carrying capacity, AI, and possible effects on smallholder farmers. I hope this explanation helps!

Thanks, very interesting. 

Further automation of agriculture could reduce the points of failure compared to an offline system.

Did you mean that this would increase the points of failure?

Thanks, was uncertain how to phrase that and evidently should've phrased it more clearly. Having lots of independently operating farms that aren't automated is more resilient (but perhaps less efficient) than relying upon a few large, highly productive, automated farms, because the failure of one has less of an impact on the whole.