Indian Agriculture is a complex sector, but for India to be a five trillion-dollar economy, it’s clear that it has to play a much larger role.

Innovation and precision technologies [e.g., IoT, sensors, weather forecasting, remote-sensing data from satellites, drones, robotics, mobile cameras, Artificial Intelligence etc.] can transform agriculture: raising yields, optimising resources, improving profitability. They yield multi-pronged impact across the complex agricultural value-chain, which is rapidly becoming digital, leading to a lot of data being generated. In addition, we now see a rapid increase in “machine-data” generated by precision farming applications: field sensors (for soil moisture etc.), satellites (for yield estimation, early pest warning etc.), robotics (for water and nutrient injection, harvesting, high precision weed removal etc.), mobile cameras (for pest attack, nutrient deficiencies), and drones (for real-time farm monitoring, surveying, 3D modelling etc.) which help farmers raise yields and optimise resources. 

This mountain of data, once anonymised, aggregated and processed can be re-purposed using AI for different use cases, provided it is interoperable and conforms to laws around Data usage. 

There is a visible need for an open, scalable, integrating platform, that democratises access to Agri information, credit, insurance and markets; incubates innovative business models; and enables better decision making. The Indian Agricultural Platform (IAP) created by the eco-system, governed by the Govt., is envisioned as an “enabling framework of Data and Services (applications) around a data exchange”. Once the IAP becomes a reality, many use cases will emerge that will seek to leverage different data sets, across entities. Agri Fintech (Credit or Insurance) or Govt decision making for Agri are two examples, very urgently needed. As an estimate, there could be 100-150 new business cases that emerge, once data collaboration and partnerships become understood and start being leveraged. 

Agricultural Data Oversight, Protection, Exploitation:

I would like to add a cautionary note, given the Kris Gopalakrishnan chaired committee report on “Non-Personal Data” which in my view, has not adequately addressed the issue of “machine-generated” data.

The very same Data-set that incubates new business models and creates additional economic value for the farmer and the nation has a flip side if misused, seemingly infringing on “personal” rights and hence subject to scrutiny. The “eye in the sky” that predicts national acreage and production by crop warns against pest attacks or verifies crop insurance claims, can just as easily help in understanding land (mis) utilisation, cropping patterns, estimate revenue (and profits) even from sub-acre sized farms. 

Hence, along with multi-stakeholder interest groups, we need deeper reflection and clarity in understanding and defining “Agri Data” within the purview of both, the Personal Data Protection legislation and Non-Personal Data (NPD) report being debated. This is especially relevant when it pertains to satellite/sensor/machine auto-generated “national or global scale” data which when processed by AI engines reveals significantly more insight than what is commonly understood even for land parcels, as small as 15m X 15m and very soon, 5m X 5m.

For instance, remote sensed data from satellites, visual data from drones, combined with AI could provide with a 80-90% accuracy, using digital signatures identify the year wise crop patterns for past several years, showing the months when the crop was planted and produce harvested, with the crop yield, hence estimating revenues. It can help show the damage caused due to cyclones and unseasonal rain or frost, predict the arrival of pest and disease a week earlier than visible to the naked eye and even count the fruit on the trees as well. Satellites can detect and delineate farm boundaries, identify water source nearby as well as detect root-level soil moisture up to 1.1 metres below the earth surface, graded every 20-30 cm or so. This capability is available at the national scale, or even at a global scale, “on tap” and can be updated every daily/weekly / monthly depending on the business case. For example, a soil moisture AI can help overcome the impact of monsoon cloud cover, thus allowing us to “see” unhindered.  

Most of all, this can be examined for sub-acre farm size, 10 M X 10 M hence there is very little “non-personal” data that is hidden.

Data usage and privacy issues stemming from the use of IoT and remote sensed data with AI should be carefully evaluated for issues like personal vs public ownership, private Vs 3rd party usage rights, processing rights, shared monetisation, inter-operability standards etc., examining these against the newly introduced terms like Data: “Controller”, “Processor”, “Principal”, “Trustee or Custodian” etc. 

We have to be particularly sensitive since much of this data pertains to Farmers who are mostly at the bottom of the financial and digital pyramid as well as to a sector like Food & Agriculture that needs critical injection of technology and innovation to drive growth and become profitable, resilient and sustainable. At the moment, farmers would happily trade data rights and anonymity for additional income. But, we have a responsibility to ensure they are not unknowingly exploited, financially bypassed by others using their farm data, for selfish business benefit and market manipulation.  

Therefore, norms for data oversight and protection have to be debated and evaluated carefully for non-PII (NPD) data as it borders on being a “public good” which can propel Agriculture forward or cripple it under the weight of the additional regulatory burden. Interestingly, the role of international satellites (Landsat, Sentinel, etc.) pertaining to similar data-sets needs discussion as well, but within the limits of the feasibility of India exercising control. It makes little sense to govern domestically sourced data in a particular way, only to find similar data-sets, available free of charge, through international Satellites for further AI processing and use. It is pertinent to point out that this is exactly what is happening at the moment since Indian laws prohibit sharing of data from IMD, weather stations, or satellites with private entities especially MNCs or with foreign investors. If made available, they charge exorbitantly (Rupees 2000/- per image in cases). Hence, start-ups actively (and freely) use international satellites sources, including IBM’s Weather Company and build business models around them. 

We should also consider an independent Data Regulator, protected by law, with suitable insertions in its charter, especially given the nuances discussed here to ensure data protection, and fair, equitable access to Agri data. Establishing good governance norms and careful deliberation on this issue of “Non-Personal Data” as pertaining to machine-generated Agri Data-sets will be crucial for the long term sustenance of this sector.

Image source: IBM, NASA Earth Observatory