Briefly describe your work with agtech and explain what motivates you to invest your time in this work.

We use the responsible innovation framework as a way to bring more inclusive tech development and evaluation through our research. I am motivated by ways in which digital technologies (esp. satellite imagery and algorithms) can be developed and used for not only predicting yield, but also for enhancing environmental quality and doing this in ways that are uplifting for the farm work and rural communities. I am also interested in seeing how humans in ag come to interact with new technologies (since agriculture is one of “our” oldest occupation).

Briefly explain any commitments to sustainability that you or your organization bring to your work on agtech innovation. Be as specific as possible regarding what kinds of social and environmental impacts you aim to produce, and the relevant strategies you are pursuing.

In our projects funded by the NSF, we are developing algorithms that are geared toward predicting Nitrogen and Phosphorus in farm use in three states: Virginia, Vermont, and South Dakota. We are also testing whether a sensor-driven, performance-based payment for ecosystem service could be “effective” to incentivize farmers to use less nutrients on their working lands.

Briefly describe the way(s) in which you assess/measure social and environmental impact in your work on agtech innovation.

In our on-the-field experiment, we collect baseline data about nutrients that farmers are using on their field. Then we use an existing process-based model (APEX) to estimate how different farming practices (e.g. cover crops) would effect the yield, Nitrogen, and Phosphorus (per lb per acre) on that field. This model uses soil test results as its input. Then we use a quasi-experiment, where some farmers are given an enhanced APEX (that uses AI) and provides bias corrected results, while others are given the current non-bias corrected APEX model results. Some farmers are given additional financial incentives (e.g. $ per lb per acre for Phosphorus reduced), while others are only given the basic incentives to participate in this quasi-experiment. Then we use difference-in-difference or some other machine learning technique to estimate whether incentives or better models have an impact on farmers’ ability to improve the environment. We then use digital serious games to test the same experiment in the lab.

Our research and this workshop aim to investigate tensions between the demands/imperatives of the tech-finance industry and the demands/imperatives of social-environmental problem solving. Please comment on this problem frame in general, and in relation to specific examples from your own experience.

Results from our research suggests that small and medium-sized family farms and ecologically diverse farms remain underserved by private sector research on Prec Ag. Private sector research is often focused on the greatest market return, i.e. decision support systems and equipment for commodity crops grown on large-scale conventional agriculture systems. These systems often are not appropriate for small and mid-sized farms attempting to mimic natural systems and grow food for local markets. Biophysical differences between large farms growing commodity crops and small farms growing food designed for local markets has resulted in technology inequity. What are some ways of designing and testing AI algorithms that can be used on small, medium, and large farms that will reduce the reliance on conventional fertilizers and improve water quality, especially when training and test datasets are limited in scope.

To investigate the tensions suggested above, we rely on the concept of “mission drift”. We understand mission drift as a tendency for social and environmental impact commitments of individuals and organizations to leak out over time due to pressures and opportunities to expand revenue, valuation and capital gains. Our project aims to investigate mission drift applied to entrepreneurial ventures as well as to organizations dedicated to supporting innovation. Please comment on this thesis in general, and in relation to specific things you have experienced where possible. To the extent you find this thesis useful, what strategies can you identify to defend against mission drift?

On our NSF funded precision agriculture / responsible innovation project, we spend a lot of time in our team discussing what “responsible innovation” meant to everyone and established specific protocols to do that in our team. That included some external measures, such as this is how we will interact with farmers, include them in the process of research etc, but also some internal measures, such as how we will collaborate with each other, the coauthoring roles of graduate students in our project, discussion about commercialization/open science, and how will some tedious tasks be shared with the team (e.g. processing multispectral imagery).

Please share something you would like to take away from the workshop.

I really want to hear from the technology developers on how they are envisioning sustainability via these new technologies in agriculture.