Agenda and Minutes

Main Topics

Use decide tag for any decisions taken

1. State of the industry with regard to data access, permissions, sharing, etc

   1. Topic Leader: @Jeremy W Wilson

   2. The desired outcome is that the group discusses where the industry is currently at for data stewardship

   3. Pre-meeting notes and references:

   4. Minutes:

      1. Who owns data?

         1. Jeremy - Farmers feel they own the data. Soil testing, it can go multiple ways. Soil samples themselves are probably the data of the farmer or landowner, but the resulting data can be owned by the service providers.

            1. Stuart - you combined access and ownership

            2. What about remote sensing data?

               1. Jeremy - I’m not as prepared to answer that question. That’s what I brought you all in for. I don’t think it changes a whole lot from how I feel about soil testing data as well. If you’re going to take UAV images… I’ve played around with it some. I want to play with some of the gamma ray sensor soil testers, because I can’t test at the frequency I need to make decisions yet. But that will likely fall like my recommendations after my soil testing results. The person that provides that service for me has something in that, and in that case we share some ownership for that data because I can’t spend that money to get.

         2. Kelly Nelson - In some of the EU ADAPT groups we talked about this, and within the EU ownership of farm data seems cut and dry. Does anyone know in the EU?

            1. Stuart - you can have one owner of a field and one owner of a car, but that data isn’t the same. It can be faithfully copied and it’s as good as the original. When you start talking about ownership it isn’t a cut and dried one owner thing. We duplicate it in a way that most things we talk about ownership of can’t be connected.

         3. Peter Schott - I don’t think the farmer owns the data. If you’re using the combine that the creator put all the money for R&D into to create, they have some claim to that data. You couldn’t plant that field without it. I signed trade agreements that they would plant but not re-plant it. Do they own the rights to the data about that seed? They don’t own the seed. There are trade offs that farmers make to license technology on their farm in certain instances. It feels good to say farmers own the data, but I don’t necessarily feel that’s right. The ownership vs how it gets used needs to induce a shift in the conversation.

         4. Seven data protection principles are

            1. Lawfulness, fairness, and transparency

            2. Purpose limitation

            3. Data minimization

            4. Accuracy

            5. Storage limitation

            6. Integrity and confidentiality

            7. Accountability

         5. What is the role of this group?

            1. Peter - who in this room is representing growers? Are there any or are we all vendors? IF so, we’re missing part of the conversation

               1. Jeremy - they are behind the ag data transparency tool as much as anybody going. There’s value in what they have done in ADTE, but in my view as Jeremy Wilson the farmer? It’s still not exactly what I need. It gives me the ability to find terms and conditions answers I need, but from my perspective? it doesn’t go far enough.

            2. Jeremy - is there a different role AgGateway can play in this and bring value to you, our members? We probably won’t define ownership. We likely won’t define access. What role can we play in this?

         6. Kelly - In my mind there is an opportunity for AgGateway to define that list of things that you stated. Simple things like was this recorded with a certain GNSS. What is the precision? Is it centimeter? Is it meter? That’s one data point. I can see us defining something like that, that menu of things that need to be defined.

         7. Has there been any work done on things like precision, results of testing, calibration? Anything to describe the data so you can go find and establish answers to this and define rudimentary data quality?

            1. Jeremy - I think 2012 was the last time we looked at that for the crop insurance industry. Our recommendation for yield monitor calibration was however the owner station could do it. We did define meta data associated with that calibration report.

               1. Stuart - I think Scott Nieman has done some work around data set meta data.

               2. Jeremy - I think it got advanced in In Field Product Identification

         8.  

      2. Who needs/wants access?

      3. How is data tracked, what access logs are needed?

      4. How is data quality determined?

         1. Kelly Nelson - At the mid year meeting we had a discussion along these lines. I think someone was looking for a data pedigree or a data manifest.

            1. Jeremy - I am responsible for John Fulton, and that was a discussion 8 years in the making. It took him four years to come to us and talk to us about it.

      5. What is the regulatory landscape?

         1. Jeremy - we talk about gdrp. There has been plenty of talk around things floating around in Washington DC about what they might do

         2. Stuart - I think the underlying principal is more of an ethics thing. With GDPR you want the confidence that the data collected is not being used against you in some way. As a result, you want the right to declare to anyone that keeping data about you that they need to delete it, even if you initially provided consent. You want the ability to revoke that consent. In the US regulatory landscape…with GDPR, I don’t know what that entails. If you’re talking about a grower applying restricted use crop protection and the government agency responsible for keeping track, the grower can’t turn around and tell them to delete all the data. The principle doesn’t apply. When you cross from data being used against you vs to hold you accountable for your public responsibility. I don’t know of anything comparable in the US.

            1. Jeremy - I have not found anything close to that.

         3. Ag Data Transparent in Canada. If there is something as an industry we could come in and support, to be a voice behind? There’s power in that. Eventually it will become a requirement to share that data. There’s power in that to help powers be more understanding when that requirement comes along.

            1. Jeremy - Is there opportunity for AgGateway to strengthen that tool that already exists? How many growers had heard of AgDataTransparent or ADTE.

            2. Kelly - it reminds me of credit card offers or bank terms. It sounds like you might want to define some sort of visual thing along those lines for growers

               1. That’s sort of what this is. Supporting that transparency.

      6. Who derives value from data?

      7. How do we handle data that is used to train AI models?Kell

2. What does the industry need?

   1. Topic Leader: @Jeremy W Wilson

   2. The desired outcome is that the group discusses what is needed to address the issue in the industry

   3. Pre-meeting notes and references:

   4. Minutes:

      1. Jeremy - What is the industry willingness to engage in this? to help build some of what we talk about? Something along this data stewardship line. We are all sharing data at some level. When I was working with EFC, we had 42 or 43 connections into other people’s software, getting some type of data. Whether that was raw data, sentinel data… We have all these connections. Is there a willingness within the industry to define this?

         1. Stuart - If your idea is moving towards a central data silo under which you can commit your data, a central data custodian and be able to control others access to it (allowing or revoking), that’s going to be a super hard sell. I’m not saying it isn’t a nice idea, it just occurs to me that it will be a super hard sell. There’s too much remaining mentality that possession of the data leads to competitive advantage, not what you can do with the data.

         2. Jeremy - Is there something our organization could do to put some guard rails in a sense around this whole data thing and get them to agree to them, to use them, to help increase growers' confidence and willingness to share data? Everyone told me that no farmers would want to share data. I have asked every grower I could find if they would share data. The three or four people I was convinced would immediately say no? Two of my neighbors said they share everything because it brought value to what they did. More growers than I can count that I thought would be willing looked at me and said “I’m not sharing anything unless someone makes me”. Is there something we can put around this we could then share out to people..there’s some of these levels of data that if we can get people to share them, we could all be better. We, as an industry, would win. As an industry, can we bring them together?

         3. Stuart - informed consent is the underlying principle. Most people just click through those consent agreements. You log onto GoToMeeting and it asks to use your microphone or your camera, you just say yes.

         4. Jeremy - When we combine the resources in ADAPT Standard and so many other standards AgGateway has facilitated, we have the ability to move our industry one step further. We can meet the demands of this new customer.

         5. It links back to the value we see in sharing that data. How do we get a message around what the value is for the agronomist or the grower in sharing data? That’s what’s missing. People don’t get the value of what they’re getting when they share their data? How do we share that message of how we will make their life easier, and minimize the risk people see in it.

            1. How do we define value in this realm?

               1. Farmers have had the value explained in their face for years, it’s a matter of getting them to actually believe it. Does it make sense, to a degree, to just sort of ask for the data without trying to define the value?

               2. Kelly - we talked about a data quality manifest, and something like a data privacy scorecard. I am particularly interested in a data quality manifest. I’m not going to touch data privacy.

               3. Jeremy - I feel like one can’t exist without the other. I don’t know if they go hand in hand, and I’m happy to stand up both. I think just taking privacy out of it and making it data stewardship would be a better choice.

               4. Kelly - I don’t know if there’s a working group there, but there’s an opportunity that’s not so controversial.

               5. Stuart - the good news is, there’s an ISO standard for that. There’s a standard that deals with data product specification. The idea is that you lay it out in a framework to express what makes data acceptable. There is a process for checking performance of data against the data product specification.

               6. Kelly - just defining what data quality is different from minimum quality data.

               7. Stuart - data quality is very subjective based on what you expect to use it for. This would be a framework for how to call out the elements that you can make pass or fail decisions about. Part of that step is identifying the important elements, which would lead to what you are interested in doing. To have confidence in a yield monitor, what would I need to know? When was it calibrated last? who did the calibration? It’s a framework for identifying those things. Then there would be a separate thing about how to execute that evaluation against a given piece of data. It has no ag slant to it whatsoever. What we would have to look into is how to apply that framework in a context of dealing with ag data.

3. What is needed to meet the needs

   1. Topic Leader: @Jeremy W Wilson

   2. The desired outcome is that the group identifies what actions if any AgGateway should take

   3. Pre-meeting notes and references:

   4. Minutes:

      1. Update the privacy whitepaper?

         1. Put together a group of people in a working group to revisit the whitepaper and decide if it needs updating or restarted?

            1. If we’re really going to care what people think? We should have USB, we should have NCGA, we should have people from those groups participating. Otherwise we should just do what we’re going to do and pass it down.

            2. Are there volunteers that would help with that? Staff can do it, but that isn’t the same as volunteers with connections that will need to be part of making that happen.

Closing Topics