As discussed in Prong 3, blockchains/DLTs are essentially platforms that help manage data, and when designed and governed using appropriate models, can help solve problems of trust, traceability and equitable data collection, storage and use. Blockchain/DLT can also support and facilitate payments (including royalty payments) of equitable compensation to those sharing their data. Blockchains can be public or private (or a combination of public and private), based on the type of architecture and governance model they adopt. Appropriately designed models can empower data providers to collectively track, control and monetize the usage of their contributed data and assets.
In particular (and summarizing Prong 3), such systems can help by (i) maintaining an immutable record of transactions/transfers of agrobiodiversity from farmers to various categories of end users; (ii) collecting data about the agrobiodiversity transferred, such as unique cultivation methods, special characteristics etc.; (iii) securing/protecting the data through the deployment of technologies such as multi-level hashing and encryption, thereby also helping ensure that the contributors and transformers of data can retain control over who can use the data, how and when it can be viewed or accessed; (iv) structuring the data to be used in the AI solution in a manageable form, and (v) facilitating automatic transfer of monetary benefits for the contributors of data, especially from their use in diverse applications, thereby effectively incentivizing continuous cultivation and engagement with agrobiodiversity.
Data managed by blockchain/DLTs, can be connected to machine-learning based AI applications to be used by farmers and researchers in search of solutions to area-specific (i.e. disciplinary or geographic area-specific) problems. What types of data can be so stored and managed by blockchains and how these can be then transformed for usage by AI solutions, needs further multi-disciplinary research from technical standpoints. For the purposes of this Position Paper and this section, we focus on ethical issues that any technical solution and legal regulation linked with such technical solutions, must bear in mind.
Ethics, Big Data & Blockchain
(Big) Data is increasingly considered a tradeable commodity, the transfer/sharing of which needs to be regulated to facilitate its beneficial use in various emerging technological applications, the most prominent of which are AI/machine learning applications. With growing economic & political relevance of data, several major ethical issues linked to its use are also surfacing. Platforms that adequately address these issues are indispensable to the meaningful & widespread adoption of (AI) applications that rely on data to design solutions for any specific sector/usecase. As discussed above, DLTs/Blockchain can potentially help address several of these ethical issues, including issues of (i) trust and privacy; (ii) secure and “controllable” data sharing; (iii) fair, inclusive and equitable economic benefits for those sharing data, and (iv) traceability (for purposes ranging from economic benefit sharing to liability determination). Yet, any blockchain/DLT solution that is adopted, can lead to additional or new problems, unless it incorporates “ethics-by-design”, i.e. it incorporates a design/architecture that takes not just existing legal and regulatory regimes into account, but also takes into account ethical concerns that are commonly linked to technological or automated solution. Some of the key ethical issues that any blockchain based solution/platform, as well as any law or policy seeking to regulate it, must bear in mind, include:
a) Trust and privacy: While AI+blockchain/DLT solutions can help address issues of trust (as discussed in Prong 3), they might also create new issues of trust, including trustworthiness of codes governing smart contracts and trustworthiness of persons and institutions running nodes on a blockchain. At the same time, privacy (and security) of those participating in the system, are a concern particularly in rural set ups where enforcement of law and order can be challenging, especially in situations where emerging technologies are viewed as likely to (at least partially) disrupt existing socio-economic power structures.
b) Fairness, bias and inclusion: As discussed in Prong 3, it is envisaged that AI+Blockchain can help overcome barriers inadvertently created by current regulatory thickets. However, as code and machine learning based systems have their own limitations and can result in the development of new types of unintended biases and exclusions, checks and balances need to be built into any AI+blockchain system aiming to equitably promote research and in situ innovation in agrobiodiversity by all stakeholders (farmers, researchers and breeders). Further, empirical research is needed to identify what is considered ‘fair and inclusive’ by contributors (farming communities), vis-à-vis the use of agrobiodiversity by downstream players and fair remuneration/royalty for accessing the same. Questions such as the period of time for which royalty must be payable following a transfer of PGRs, the % of royalty, the means by which this royalty is to be utilized (e.g. will it be only for individual farmers or for farmer communities, and the funds can therefore be used for community development), are all questions that relate to the ethical issues of fairness, bias and inclusion. Inclusion vis-à-vis such technologies is also a matter of making sure that equitable access to hardware (smart phones) and internet is available to small farmers, not just in specific regions of India, but across all regions. This would need, for example, the focus of subsidies to shift or re-distributed to cover not just fertilizer subsidies, but subsidies for acquiring the necessary hardware and internet access.
c) Transparency (including explainability) and traceability: Transparency vis-à-vis sources from which data is collected and the end uses to which it is put, is crucial to building trust in the system and ensuring its long-term usability. While blockchain/DLT solutions enhance transparency and traceability to source vis-à-vis digital data, they are only starting to be appended to AI and IoT devices to permit traceability also of physical goods. In case of seeds and soil microbial diversity, which are sourced with the aim of further transformations, in order to make the traceability meaningful, they would need to be combined with biomarkers (or similar) technologies. Ethical and multi-disciplinary issues linked to such technological combinations will need to be investigated.
d) Governance, regulation & sustainability: With AI & blockchain technology replacing human actors, it would be necessary to ensure smooth interaction between existing governance structures & regulations on the one hand, and emerging AI/blockchain based technological solutions on the other, to ensure a sustainable and seamless transition that maintains and secures meaningful and continuing interaction between human and autonomous actors. In this situation, one of the key questions that would emerge would be: What kinds of organizational and leadership models/structures can support synergistic interactions between human actors (in current regulatory regimes), and autonomous actors (codes) in planned AI+blockchain applications to enhance trust, promote equitable benefit sharing and ensuring responsible decision making? From a more practical standpoint, it would also be necessary to determine which government agencies, NGOs and private players would need to be ‘nodes’ in the blockchan and whether public permissioned / public permission less (or other) architectures would be better suited to enhance trust and secure privacy in the AI+blockchain facilitated system adopted for the purpose of promoting sustainable seed innovations? With the emergence of “code” based governance, it is also necessary to see how issues of liability would be reconciled. Finally, in a country as diverse as India, fair and inclusive DLT/blockchain governance models must take cultural diversity, equity, and practical usability into account, supporting the development of ethical business models for the benefit of farmers, researchers and the environment.
Of course, the benefit of any distributed DLT or Blockchain technology increases with the number of players (nodes) and contributors (users); the greater the number of nodes, i.e. those who are engaged in contributing or testing seeds, soils, cultivation methods etc. on a blockchain, the higher the chances that any user of the system will be able to get an accurate view of the quality of the products being offered via the blockchain facilitated marketplace. In order to ensure that the system is not overtaken by vested interests also, a large number of players (farmers, researchers, end consumers, government bodies etc.) must be a part of the blockchain network.
Annex 3 in the forthcoming position paper, suggests a very preliminary outline for a blockchain/AI based framework to promote sustainable seed innovations. However, extensive multi-disciplinary research and multi-stakeholder consultations are needed to build up to a complete workable architecture before it can be rolled out. What must be underscored, however, is that emerging technologies such as Artificial Intelligence and Blockchain can and should be given adequate research as well as policy level/regulatory attention. Also, to the extent that these technologies aim to bring benefit to the poorest and most marginalized segments of society and hold the potential to incentivize research and innovation in thus far neglected areas, it may be more beneficial to invest in researching these technologies thoroughly and deploy them under comprehensive legal and ethical rules, rather than rejecting or limiting their potential scope of application and utility.
It is noteworthy in this context that while blockchain technology can support private ordering and self-governance by the blockchain ecosystem, in fields as sensitive and important as agriculture, blockchain codes and codes governing smart contracts should ideally not be privately ordered. However, semi-private ordering of codes, after consulting farmers, NGOs, scientists and government agencies would likely be more beneficial. This can entail the creation of ethical codes via multi-disciplinary research engaging all stakeholders in consultations; or self-regulation by farmers supported by broad legislative guidelines and regulatory check posts (e.g. mandatory government body or ‘watch dog’ nodes in any blockchain architecture created for promoting sustainable seed innovations).
 Russell, Dewey, and Tegmark, (2015); Fatima and Pasha, (2017); Jordan and Mitchell, (2015).
 Kamath, (2018); Imeri and Khadraoui; Agrawal, Sharma, and Kumar, (2018); Lin et al., (2018); Xu et al., (2019).
 For further details, including a better explanation of why multiple players make a blockchain/DLT solution better, please see Annex 4 below
 Kalkanci, Rahmani, and Toktay, (2018); Thomason et al., (2018); Bartoletti et al.
 Kochupillai, (2019a); Hericko; Ribitzky et al., (2018); Beck and Müller-Bloch, (2017); Iansiti and Lakhani, (2017).