Revealed: Ministry of Agriculture’s MOU with Microsoft on the Agristack. Urgent need for transparency and consultation!

Tldr

On 19th May, 2021, in response to an RTI query by us, the Ministry of Agriculture provided us a Memorandum of Understanding between the Department of Agriculture and Microsoft, and a document titled “Standard Operating Procedure (SOP) for verifying Farmer’s Database (100 Villages)”. The MoU appears to propose that farmer’s data will be used to authenticate their identity for the receipt of entitlements under agricultural schemes. Further, the MoU speaks of a data sharing agreement, wherein the Agriculture Ministry will share farmers’ data with Microsoft. While the SOP contains guidelines for correcting gaps and errors in land records concerns on data protection, exclusion to entitlements and algorithmic governance emerge.

Background

Recently, the Ministry of Agriculture signed a Memorandum of Understanding (MoU) with Microsoft for a pilot project in 100 villages in 6 states. Under this MoU, Microsoft has been roped in to

"to develop farmer interface for smart and well-organized agriculture, including post-harvest management and distribution".

The press release and subsequent media reports have indicated that a central farmer database would be prepared by linking the land records of farmers across the country that would integrate data from PM Kisan, Soil Health Cards, and the Pradhan Mantri Crop Insurance Scheme as well as other data. Geo-tagging of agricultural assets shall also take place.

This database will be leveraged by Microsoft and its local partner CropData to build post harvest management solutions and capture agricultural datasets such as crop yields, weather data, market demand and prices. This pilot sounds very similar to the "Agristack", which is the collection of technologies and digital databases, proposed by the Central Government focusing on India’s farmers and the agricultural sector (for an explainer on the Agristack, please see here).

A joint letter for transparency and consultation

In response to this, on 5th May, 55 organisations (including IFF and led by various farmer groups and agricultural cooperatives) had written a joint letter to the Union Minister for Agriculture and Farmers’ Welfare. The letter highlighted a number of key concerns with the proposed project such as usurious lending practices, exploitation of farmers by procurers, reduced agency on the part of farmers, reduced transparency through algorithmic governance, and inadequate and incorrect land records. Additionally, the letter stated that:

  1. Extensive consultations with farmers: The Ministry should hold consultations with all stakeholders, especially farmers’ organisations, on the direction of its digital push as well as the basis of partnerships, and put out a policy document in this regard after giving due consideration to feedback from farmers and farmer organisations. As agriculture is a state subject, the central government should consult the state governments also.
  2. Need for adequate legal framework: All initiatives that the government has begun with private entities to integrate and / or share multiple databases with private/personal information about individual farmers or their farms be put on hold till (i) the policy framework mentioned in the earlier point is finalised by the ministry and (ii) a data protection law is passed by the Parliament. This agreement and any other such agreements with any IT organisations must be backed by adequate legal systems in India. Existing laws must be reviewed considering the advancements in the field of artificial intelligence, machine learning and big data and new laws need to be enacted accordingly.
  3. Transparency and financial disclosures: The MoU signed between the Ministry of Agriculture and Microsoft India on 13th April, 2021 should be put in the public domain on the website of the ministry, apart from being put on hold as per Point 2 above. Further, any financial commitments or implications must also be put on hold and be disclosed.

Transparency is an important part of our work, and so we had taken a proactive position with respect to the third demand: on 23rd April, 2021, we filed an RTI request before the Ministry of Agriculture (Agri Ministry) and Farmers Welfare. Two days ago, on 18th May, we finally received a response to our RTI, in which the Ministry furnished us with a copy of a MoU dated April 13, 2021, between the Department of Agriculture, Cooperation and Farmers Welfare at the Agri Ministry and Microsoft Corporation (India) Private Limited.  We were also provided with a document titled “Standard Operating Procedure (SOP) for verifying Farmer’s Database (100 Villages)”. We are thankful to the Ministry for Agriculture for providing us this information and urge them to resort for proactive disclosures on significant policy issues that help improve citizen feedback and trust in policy formation.

But, what is an MOU? Concerns emerge.

Before we discuss the MoU, we ought to set out that a memorandum of understanding is a non-binding document, it is not enforceable in a court of law, and it ordinarily has very bare, broad and illustrative provisions. Typically an MoU is executed only to set out, in writing, what the understanding is between two parties, and to provide a roadmap for negotiations pursuant to which a binding contract is executed between those parties. However, since the MoU is a roadmap for negotiations towards a binding agreement, and also sets out the understanding between two parties, it is extremely important and helpful for us to understand some of the conversations around AgriStack and data protection.

We pored over the MoU and found some interesting points. Exhibit A of the MoU (page 7) has a summary of the project and bullet points on what the government will contribute and what Microsoft will contribute. Broadly, the solution proposed is that farmer’s data will be used to authenticate their identity for the receipt of entitlements under agricultural schemes.

It is absolutely essential here to note that under the MoU, ownership of the data, and the responsibility for its security will reside with the Ministry of Agriculture. However, the Ministry will share its datasets with Microsoft. Absolutely no clarity has been provided on how the data will be protected.

The MoU (on page 7) also requires the States to participate in this project to integrate the national farmer database and build a unified farmer service interface. However, the State departments have not been made parties to this MoU, even though agriculture is a state subject under the Constitution of India.

Further, the Non-Disclosure Agreement (NDA) at Appendix A on page 10, appears vague on the subject of whether farmers’ data shared under the MoU is considered “confidential information”, and thus subject to protection under the NDA.

The SOP for verifying the farmer database

As stated in the SOP, the Department of Agriculture and Farmers’ Welfare intends to create a farmers database and) a framework for a digital ecosystem for Agriculture. The SOP further states that since farmer data is available under various central and state agricultural schemes, a list of farmers shall be prepared from the scheme data and compared with land records data.

Now, it is well known that land records in India, especially records for rural areas, are riddled with errors and inconsistencies. This has been tacitly accepted by the SOP which states that even though some amount of matching was achieved, significant errors and mismatches remain. To resolve these problems, field visits to the 100 selected villages will be conducted under the Department of Agriculture. Each field team shall be provided with pre-filled excel sheets, on the basis of which they verify and correct the relevant details, including the name of the farmer, their Aadhaar number, the area of the land holding, and the schemes from which the farmer has benefitted.

Data from a 2017 PRS Legislative Research report on ‘Land Records and Titles in India’ suggests that the aforementioned issues may not have been resolved easily, as indicated by the slow progress of the Digital India Land Records Modernisation Programme :

  • 86% of all land records have been computerised
  • However, only 46% of cadastral maps have been digitised
  • Only 39% of records have had their spatial data verified
  • Only 28% of records have been issued a digitally signed Record of Rights
  • Only 15% of land records have real time updation of the Record of Rights and the accompanying cadastral maps
  • Only 35% of the total area under inquiry has been surveyedOnly 9% of villages have seen survey work completed

Thus, addressing these issues will require time and patience. Most importantly, the policy design should be based on studies, feedback and consultation with farmers.

Can databases resolve land disputes in rural india?

The raison d’etre of the field visit is to identify the gaps and inconsistencies within the data and reconcile them. However, therein lies the rub: what happens in case there is disputed land? The SOP provides for handling issues related to unrecorded land, new farmers who are not present on land holding lists, and land holdings whose location has been incorrectly recorded; however, no clear process has been provided to deal with disputed land. Instead, the SOP simply states that the data collected must be “reconciled 100%”.

Resolving land disputes is a key part of cleaning up India’s land records database: a World Bank study estimated that about two-thirds of all pending cases in the country are land related disputes. Meanwhile, the NITI Aayog has estimated that “on an average it takes twenty years for a real estate or land dispute to be resolved”. Given the pervasiveness of land grabs (which have been increasing in recent years) and encroachments, such issues may not be easily resolved, and will likely require a certain amount of time to forensically analyse each case to ensure that large scale exploitation and dispossession does not take place.

A GRAIN report titled ‘Digital fences: the financial enclosure of farmlands in South America’ has already brought to attention how digitalisation has enabled the land grabs by large scale agribusinesses in Latin America, while the requirement of digital land records for access to public schemes for agriculture has resulted in large scale exclusion. Thus, here, the Agristack may be a double-edged sword: in case such cases are rushed over and dealt with in a perfunctory manner, the resultant injustice in the form of dispossession would be locked in, whereas if land disputes are adequately addressed and resolved, India’s farmers could rest assured that their land holdings are have been digitally enshrined. Thus, there is an urgent need to address all such land disputes before creating such a database, especially since compensation-based dispute settlements have generally led to highly inadequate compensation for farmers and lengthy delays in payment.

Lastly, the project fails to mention landless or tenant farmers. Marginal farmers i.e. those with less than 1 hectare of land, constitute 68.5% of all land holdings as per the Agricultural Census of 2015-16. Meanwhile, as per the National Census 2011, 14.43 crore people are landless agricultural labourers. In terms of households, the Socio-Economic and Caste census tells us that 56.41% of rural households (10.1 crore households) do not have land.

These are astounding numbers, and so it is imperative that any support provided under agricultural policies address the landless. However, prima facie, the proposed database excludes landless farmers from its ambit, and so any potential benefits that are forthcoming will not be enjoyed by these households. Linking the identity of a farmer with land ownership brings other issues: for example, many estimates suggest that women constitute 60-70% of all farmers, but since they rarely own land, most schemes and technological inputs would not be made available to them. Additionally, tenant farmers and sharecroppers, who also constitute a large section of all farmers, would also not be recorded in such a database.

What Next?

We will continue to monitor the status of the Agristack project, trying to ensure that the government engages with farmers and listens to their needs instead of experimentation that can lead to faulty policy design and results in digital systems that result in harm. In the meantime, we are planning to hold a conference on the Agristack to bring together stakeholders from farmers groups, kisan unions, agricultural policy organisations, and civil society. Through this conference, we hope to grow public awareness about the Agristack and learn more about its potential benefits and drawbacks. We will be making a public announcement on this soon with other organisational partners to initiate greater conversation around this significant digital policy that will impact the rights of millions of India’s farmers.

Important Documents

  1. Copy of the MoU dated April 13, 2021, between the Department of Agriculture, Cooperation and Farmers Welfare and Microsoft Corporation (India) Private Limited (link)
  2. Copy of the document titled “Standard Operating Procedure (SOP) for verifying Farmer’s Database (100 Villages)” (link)
  3. Copy of RTI response from Ministry of Agriculture and Farmers Welfare enclosing the MoU and SOP (link)
  4. Previous blogpost dated 4th December, 2020 titled “The AgriStack: A Primer #SaveOurPrivacy” (link)
  5. Previous blogpost dated 6th May, 2021 titled “#StandWithFarmers: Joint Letter by 55 Groups | Hold the Agriculture Ministry and Microsoft India Accountable” (link)