Did you know that the digitalization of agriculture could affect farmers’ rights?

Submitted by Laura Dunn on Thu, 2021-12-09 12:07

Digital agriculture, as being proposed by giant agribusiness and data companies, should come with a warning for farmers, farm workers and food vendors. Read our blog below to find out about why in more detail (with examples) and about alternative approaches (also with some real world, feasible examples). But if you are short of time, check out our bullet point ‘what to watch out for’ summary:

Farmers:

Giving data away to large companies enriches them and makes them more powerful, further skewing the balance of power in the Industrial Food Chain
Signing up with a digital ag company could lock you into a contract that will force you to buy the products they promote and sell on credit (at high prices), follow chatbot advice to qualify for insurance (which you must pay for), and receive payment via a digital money app (for which there is a fee).
If you make any mistakes or do not comply, it could potentially affect your credit worthiness and future access to finance and markets.
Even if a company declares it will share some of the risk with you, you may find yourself locked into selling your farm products to just one company at a price determined by their algorithm (potentially on the basis of unknown criteria).
You may be forbidden from repairing the equipment on your farm.
You may find land rents and prices increasing as data companies identify and micro-target the most productive land and buy it up.

Food workers:

You may find your jobs replaced by machines such as robots and drones.
You may find that you are expected to work at the same pace as a robot, perhaps by wearing robotic devices, even though this has already been leading to more accidents.

Food vendors:

You may find that giant data platforms skew food distribution routes, inserting themselves as middlemen between farmers and consumers, and excluding food vendors (this has already happened around Nairobi).

***

What is digitalization in the food system?

In the last few years, food corporations have been getting bigger and fewer. At ETC Group, we’ve been tracking this trend of mergers and consolidation in the food system. At the moment, four corporations control over half of the commercial seed market and two thirds of the agrichemical market. These companies are Bayer (which bought up Monsanto), Corteva Agriscience, Syngenta and BASF. These mergers will continue and we’re likely to see new mergers in the future. Similar trends are occurring in other sectors like farm machinery, where the top 6 farm equipment companies (including John Deere) control 52% of the global market.

However, there is a new trend developing in addition to this: the digitalization of agriculture. The same agri-giants that grew rich and powerful from the industrial food system are now combining their power with that of data giants. The overall logic is towards an integration between the companies that supply products to farmers (pesticides, tractors, drones, etc.) and those that control the flow of data.

When wielded by technology firms, data can be an extremely valuable resource. Although many of us give away our data without even realizing it, it increases in value as more of it is hoarded, aggregated and analysed. Thus digital farming involves farmers handing over data to corporations to analyse. This is presented to farmers as a way of making their farms more efficient so that they can make a greater return on their investments. But while farmers may think they are receiving a service from companies like Bayer, these companies are actually mining farmer’s data for free and hoarding it to increase their power.

How does Digital Farming work?

One example of digital farming is the data platform Climate Fieldview which is owned by Bayer. It was developed by The Climate Corporation, a Silicon Valley start-up bought up by Monsanto in 2013 for around $1 billion [1]. At the time, people were shocked that Monsanto, a seed, pesticide and agricultural biotechnology company, would buy a digital platform about weather information – what was the relationship? Well, it seems that someone from Monsanto had the foresight to imagine that in a few years, Monsanto might decide to become a technology information company collecting and using agricultural data.

Today, Climate Fieldview claims [2] to collect data that covers 60 million hectares in 23 countries. It offers specific services like providing information about the state of a farmer’s soil. It also had a feature called Xarvio SCOUTING [3] where a farmer could take pictures of diseases on their crops and then upload these pictures onto a cloud where they were compared with some 100,000 photos Bayer already had. The platform could then suggest to the farmer which Bayer products and herbicides they should use against the disease. Xarvio was acquired by BASF in 2018.

There are already many similar data platforms, owned by big agricultural giants like Corteva, BASF, Yara and John Deere. They extract data on seeds, soil, fertilizer and weather for their Artificial Intelligence (AI) systems to use, and in turn they suggest what and when farmers should farm and what products they should use. In the developing world, these services – especially those from startup companies – are marketed as being free to use. Nevertheless, data is being collected from farmers for free and used by the companies, almost like another giant commercial harvest that they can profit from. But there’s even more to it than that – the digitalization of agriculture threatens to bring many more problems for farmers as we shall see.

Which digital farming platforms are already in Africa?

Microsoft’s Azure Farmbeats

Over the past two decades tech billionaire Bill Gates has poured a huge amount of money into getting small farmers in the Global South to adopt what he calls the most highly advanced seeds, pesticides and fertilisers [4], sold by the world's largest agribusiness corporations through programmes like the Alliance for a Green Revolution in Africa (AGRA). However, adoption rates for these technologies has remained low so far, so Gates has been banking on digital agriculture to help.

In September 2020, Microsoft and AGRA began to expand Microsoft’s digital platform Azure FarmBeats with a "chatbot" app called Kuzabot [5]. This app provides small farmers with advice via WhatsApp and SMS, including information on what inputs to use and which companies to buy from. At the same time, Microsoft's FarmBeats is integrating the US-based start-up Climate Edge into its platform. Climate Edge aggregates data on small farmers supplied by agricultural consultants, NGOs, companies, and researchers and then sells this information onwards to insurance companies, pesticide dealers, and large food companies like Unilever and even NGOs [6].

Arifu

The leading provider of "chatbot" advisory services to small farmers in Kenya, Arifu, is partnered with multinational seeds and pesticides company Syngenta. Arifu essentially "creates a demand for Syngenta seeds.” [7] It is now part of a larger digital platform in Kenya called Digifarm, operated by the mobile network operator Safaricom (which is owned by UK company Vodaphone). Digifarm provides millions of small farmers in Kenya with chatbot services like Arifu, sells them inputs and crop insurance, offers them loans, and buys and sells their produce, all via M-PESA, a mobile money transfer service. For this, Safaricom charges a fee on all transactions.

Digifarm, and similar platforms in other parts of the world, say they are providing financial services to rural people who would not otherwise be able to access them. However, this obscures what is really happening. Farmers must conform to this system: they must buy the inputs that are promoted and sold on credit (at high interest rates); follow the "advice" of the chatbot to qualify for crop insurance (which they must pay for); sell their crops to the company (at a non-negotiable price); and receive payments on a digital money app (for which there is a fee). Any missteps can affect a farmer's credit worthiness and access to finance and markets. According to the organization GRAIN, this is essentially “contract farming on a mass scale.” [8]

What about automation? Is that also part of digital farming?

Yes. The vision of digital farming being pushed onto farmers includes automation. This includes the use of drones. Agriculture is set to become the sector with the second largest use of drones, both for scouting the farm and capturing images and for the application of pesticides. Automation will also involve a massive uptake of robots, especially mini-robots that can weed, pick berries and deal with pests.

The overall idea of digital farming is to bring many of these processes and technologies together: drones, extraction of data, robots, smart farming systems, etc. It essentially comes down to creating a food system stripped of all direct human relations with the soil, plants and living matter, mediated by data-driven corporate strategies.

What are some other possible impacts of digital farming?

Some of the main concerns and issues with digital farming are the following:

A lack of control and ownership

When farmers buy equipment like tractors from John Deere, they must sign up for certain conditions with respect to the data that John Deere will gather from the tractor. In addition, farmers are sometimes forbidden from repairing their own equipment. This raises a number of concerns about losing control and ownership over both data and the tools used in farming.

Destroying the livelihoods of food vendors

These new agricultural technologies are being introduced within an ideological “framing” advanced by powerful companies, which promotes their commercial interests. Some of them start out with the claim or the intention of benefitting small farmers and food sellers.

One example is Twiga Foods in Kenya, an agri-tech start-up company founded by a US academic who wanted to connect farmers to small vendors, bypassing powerful food cartels. Twiga Foods built up a fleet of trucks to source foods from farmers outside of Nairobi, and then deliver this food directly to a network of small vendors in the city. All of the transactions were organised through cell phones and run on Microsoft's digital platform and Azure’s cloud services. Twiga's early success caught the attention of larger companies. Goldman Sachs and the French family that owns the Auchan supermarket chain took major stakes in the company and then Twiga partnered with IBM, another top cloud services provider, to pilot a digital banking scheme with its vendors. Twiga is now selling foods directly to consumers – cutting out the small vendors it was initially established to serve. Twiga's more significant impact is that it has refashioned food distribution, using pretty much the same workforce, to enable corporations to insert themselves in the middle and extract wealth.

Pushing up cost of land

In 2020, the Bayer Fieldview app signed an agreement with Tillable, a company which calls itself the “Airbnb of farmland.” They are building a database of farmland which can be rented out to farmers. After this agreement, it seems that some tenant farmers who used Fieldview began to receive demands from their landlords to pay higher land rent. It appeared that Tillable had taken Bayer’s Fieldview production data and used it to revalue land, and had then approached owners to acquire the land leases, renting the land out at higher rates.

Selling data: to whom?

When data is aggregated from many farmers, it no longer belongs to them and it can be sold on to third parties. Climate Fieldview, the platform now owned by Monsanto, has made agreements with drone companies, farm machinery companies like John Deere, and weather companies to share and use the data that they collect. While it sounds as if this model of digital agriculture might help farmers, the data that is being extracted from them is extremely valuable and it is unclear who it is being shared with and sold to, and what it is being used for and to what end. Many technology companies are in turn owned by shadowy asset management companies like Vanguard and Blackrock. It is unclear if they are also getting this data from farmers (in an aggregated form) and if so, what impact this will have on small farmers.

Lock-ins

In some instances a farmer can sign up to Bayer’s Fieldview app for close to free, and in return, they are promised certain returns. For instance, if they use Fieldview’s prescriptions about how and what to farm, they may be promised that they will be able to sell their produce at a certain rate. Bayer is experimenting with ‘outcome-based pricing’ plans: if the farmer is not able to sell the produce, Bayer will make up the difference, but if they do sell the produce for more than the rate prescribed by the app, Bayer will receive more than half of the difference. The farmer has no insight into how these calculations are made or whether they are fair – they are created by an algorithm, which is so hard to understand that it acts like a “black box.” Bayer claims that this model works for farmers since they are sharing the risk with the farmer, but what it really means is that farmers are both passing on their data to the company and agreeing to buy their products, locking them into a relationship with the company which the company will use to extract ever more value from them.

Robots versus farm workers

The agricultural industry is notorious for exploiting workers and treating people as disposable labour. This is evident in the case of seasonal migrant workers in North America for example. With the introduction of robots into digital farming, there is even less need to hire skilled workers and migrant labour. In some cases, robots are being used to augment or enhance the labour of workers. In Japan for instance, pickers wear automated arms to increase their ability and the speed with which they can pick things up. Roboticizing workers has led to serious bodily injuries. One of the industries that has been highly robotized is Amazon’s warehouses; workers are more controlled, they are directed as robots would be and they are expected to process items faster and faster – again, leading to injuries. An investigation into worker injuries in warehouses which had robots found that injuries were 50% higher than in warehouses that did not use robots [9].

Money in speculative assets

Digital agriculture companies claim that they can reduce climate-damaging carbon emissions and create carbon sequestration in the soil by using sensors and providing automated prescriptions for climate-friendly farming practices.

They then collect the data from these sensors, or the promise that farmers are following certain practices, to generate “carbon credits” that can be sold to carbon markets. Bayer for instance has its own carbon trading platform. A large amount of profit lies in these speculative assets, through carbon credits being sold from digital farming. There is now concern that these techniques may not ultimately store much carbon, but carbon credits are still being bought by polluting industries, giving them a free pass on the basis that they are “offsetting” their carbon emissions. Furthermore, in the process of generating carbon credits these systems do not count the carbon produced for energy to power digital processing, movement and storage of data.

A solution to climate change? Not quite

It is easy to believe that digitalization will lead to lower emissions and therefore to combating climate change. But the opposite may be true – data may in fact be part of creating rather than reducing climate emissions. Data is not intangible: it is energy, and it requires a huge amount of infrastructure including wires, cables etc., which in turn rely on energy and involve water consumption, chemical waste production and the mining of minerals. Digitalization will generate a huge amount of data which has to be stored and processed: by 2025, it is estimated that a fifth of global electricity consumption will be taken up by data – that is equivalent to 14% of global emissions. Artificial Intelligence, likewise, requires a huge amount of energy. Training one AI system can be the equivalent of traveling 700,000 km by car [10].

What if the whole system gets hacked or the internet goes down?

If the entire food production system depends on data and data networks, what happens to food security if it goes wrong or if the data networks fail? We do not know why an AI makes the decisions it makes. A typical example of the dangers of this unknowability is when a self-driving Uber car dependent on AI killed a 49-year-old woman pushing a bicycle because the algorithm did not see or recognize her [11]. Uber, the company who owned the car, was found not responsible for her death because it was a computing mistake. So if the entire food system becomes dependent on a system reliant on AI and it is not clear exactly how it works, who will be responsible when something goes wrong? Making food dependent on digital infrastructure also makes it a target for cyber attacks.

Can digital farming be done differently to be beneficial to farmers? Are people fighting back?

For digital farming to be in any way beneficial for farmers, they would have to take back control over their data from giant tech and agriculture firms and acquire the capacity to benefit from it by also having the capacity for high-power data processing. Publicly-generated environmental data, genetic data, weather data and agronomic data must, at the very least, remain in the public sphere, free from commercial exploitation.

Some initiatives offer free and open source technology that is not controlled by those who own patents or gather data. One example is the Farm Hack Initiative, a global collaborative platform for sharing farm tools and knowledge among farmers from across the world. Farm Hack is part of the Right to Repair movement, which is an important part of questioning the role of data in food systems and asserting people’s rights and control over data and data-driven technologies.

Farmers and communities would also need to expose and challenge the attractive deals and free apps extended by technology companies keen to suck up knowledge and data to improve their algorithms and machine learning capacities. These companies must be required to obtain the Free Prior and Informed Consent of farmers and communities before collecting data from their fields and agricultural practices.

What other alternatives exist to digital agriculture?

Agroecology, also known as ecologically-just forms of ecological food production, builds on existing practices by smallholders and peasant farmers and is the best way forward for creating robust food systems that benefit farmers, small vendors and consumers. Several groups across the African continent including the Alliance for Food Sovereignty in Africa (AFSA) and COPAGEN are advocating for agroecology.

Mutual aid networks can connect smallholder food producers with local consumers, in the midst of disruptions in export markets and commercial supply chains, e.g. during lockdowns. For example, in Karnataka in India, farmers started using Twitter to post videos of their produce and connect with buyers during the COVID pandemic. This can re-create mutual aid networks between producers and consumers, removing the powerful cartels that have come in between.

Technology assessment (TA) is fundamental to provoking debate and ideas about what could be a fair, just and ecologically sustainable use of digital technologies that serve the common good. Participatory Technology Assessment is a process that enables civil society, Indigenous Peoples, local communities, farmers, fisherfolk, and popular and social movements to come together and evaluate new and emerging technologies as well as examining the interests and powers behind the introduction of new technologies, the ways in which those technologies are applied, and their potential impacts on the environment and communities.