Dr. Éliane Ubalijoro, Chief Executive Officer, CIFOR-ICRAF, Kenya

Abstract

As the largest source of employment and a major source of greenhouse gas emissions, the global agriculture sector has enormous potential for helping alleviate climate change, poverty, food insecurity, drought and other emerging and entrenched challenges. Trees are a cornerstone of any viable pathway to reducing the negative impacts and magnifying the positive impacts of the agriculture sector. At the intersection of trees and agriculture, emerging technologies and the data underlying them are supporting the wider uptake of sustainable and regenerative practices. There is an opportunity for significant developmental progress against global inequalities through the application of emerging technologies in agroforestry. With productive climate resilient agriculture that promotes biodiversity conservation, landscape health can be assured from farm to forests, decreasing pressures for land-use change while promoting corridors for nature to thrive. To achieve such goals, and to develop technologies that are fair, ethical and recognize the rights of all stakeholders, it is important to develop a data governance framework that optimizes inclusivity, accessibility, interoperability and reliability. This paper explores principles for a digital commons supportive of emerging technologies at the intersection of agroforestry, ecosystems and human health, seeking a framework that maximizes the potential of data-centric technologies while leaving no one behind. Properly conceived and implemented, a digital commons can support the systemic approach of the Planetary Commons, thereby enabling greater planetary resilience.

Introduction

In March 2023, the Future of Life Institute published a document entitled “Pause Giant AI Experiments: An Open Letter”. The letter was signed by over 30,000 people, including eminent technology pioneers, and was a response to the wide and growing concerns over our inability to control the pace and direction of AI. Despite widespread support from prominent individuals in academia, philanthropy and the private sector, the letter failed to achieve its goal of a pause, and the development of AI has only increased since the letter was published. The letter stands as a cautionary example that support from individuals and organizations may not be sufficient to affect change in the modern course of technological development. The economic and political forces that set such a course are powerful and require new approaches if they are to be moderated in the interest of ensuring that benefits accrue to all people rather than a select few.

A part of the modern digital economy that has grown exponentially in importance is the data and data infrastructure that forms a necessary component of AI systems. Such data pertains to a wide variety of subjects, from people and places to processes and actions. An unfortunately common thread is that the varying stakeholders are struggling to agree and implement principles for governing the processing of data, even while technology continues to advance into new areas and applications.

Notwithstanding the challenges, there is an opportunity for data to be a source of strength and unification. The collection of datasets opens new possibilities for applications that strengthen our collective resilience in the face of global crises. As planetary and local environments change, efforts towards adaptation and mitigation will increasingly require access to such datasets; limiting access only hinders our development and ultimate success in the face of monumental challenges. At the same time, unlimited and unregulated access provides too many opportunities for uses that are contrary to the interests and wishes of data subjects, contrary to principles of fairness, and contrary to long-term development of communities.

This paper explores principles for a digital commons[1] supportive of emerging technologies at the intersection of agroforestry,[2] ecosystems and human health, seeking a framework that maximizes the potential of AI, quantum computing and other data-centric technologies while leaving no one behind. The paper concludes with recommendations for approaching the development of such principles, particularly considering the failure of past attempts to slow the uncontrolled progress of emerging technologies.

2. Contextualizing data and agriculture in the Anthropocene

The anthropogenic era started a few hundred[3] or a few thousand years ago,[4] and is characterized by anthropogenic emissions of greenhouse gases. Regardless of the starting point, the anthropogenic era is significantly tied to environmental changes from agricultural practices. Whether deforestation occurs to clear land for farming or to support industrial activities, it is accompanied by significant negative impacts to soil and water, loss of biodiversity, emissions of greenhouse gases, and an overall loss of resilience.

As a major driver of climate change and other environmental impacts, the agriculture sector is also key to addressing these challenges. An agriculture sector focused on sustainability and environmental restoration holds the power to enhance soil health, conserve water and increase biodiversity. To achieve these outcomes, it has been suggested that humanity must free itself from “dominating egoism towards nature” and replace capital-oriented decisions with transformational policy in terms of agricultural resilience.[5]

Within the last decade, the agriculture sector has adopted modern technologies at a rapid pace, and that pace is accelerating. With the introduction of mobile-based apps, smart agriculture and now artificial intelligence (AI), the global agricultural industry is highly industrialized and data-driven and, as a result, highly influenced by those who control data infrastructure. Consolidated control over the means and expense of collecting, managing and processing relevant datasets can exacerbate inequities in the agriculture sector, with significant impacts on food security, poverty, climate change and planetary resilience. Efforts to counter these inequities are consistent with the Sustainable Development Goals and the Montreal Statement on Sustainability in the Digital Age.[6] One avenue for such efforts seeks solutions through data governance that is inclusive, responsive and responsible.

Certain concepts and principles of data governance are older than the digital age but grew in importance with the development of computer networks. In connection with the growth of AI and other emerging technologies, data governance is now a fully-fledged academic sub-discipline, incorporating aspects of law, technology, ethics and philosophy. Attempts at reducing data to a mere commodity (“data is the new oil”) provide a convenient economic perspective but ignore or severely discount the personal nature of data.

The concept of a commons is an ancient one that is integral to the concepts of civilization and community. Like data governance, the commons was of interest in the digital age prior to computer networks, but interest has grown significantly with the development of data-centric emerging technologies. Notably, while many technology companies are willing to provide AI algorithms and tools to the public through open-source licensing schemes, those same companies are far less generous with the datasets they build to support AI systems. Demonstrating the critical importance of data, technology companies go to great lengths to obtain, maintain and restrict access to the largest possible datasets.[7]

The application of principles of a commons to the development of technology is not new. Numerous examples have demonstrated the feasibility of a commons, including GitHub and Creative Commons in software development and digital works of creativity, respectively. In the context of data and information, there is yet to emerge a commons with depth or scope equivalent to GitHub or Creative Commons. The economic value and current proprietary nature of data are not the only challenges preventing such an emergence. As described in the next section, data may exist in many formats, is often highly contextualized, may be significantly time sensitive, may be highly biased or non-representative, or may violate data protection or copyright law, among numerous other possible issues. Notwithstanding these challenges, success in building a digital commons will compound successes in the development of technologies that work in service to people and the planet.

3. Bridging science, agriculture and tradition for the challenges ahead: Building the digital commons

Data pertaining to trees and agriculture have many sources and formats and are highly contextual. Datasets vary in characteristics that include resolution, geographic coverage, frequency of measurement, consistency of measurement protocols and other factors. Often, higher quality datasets are proprietary, leading to inequality between users with and without access. Nevertheless, citizen science and the proliferation of mobile applications and smartphones has drastically increased the availability and quantity of agricultural data. A few of the sources of agricultural data are discussed in the following paragraphs.

Earth observation (EO) data from satellites cover the ultraviolet, visual, infrared and microwave bands, and allow time-dependent mapping and monitoring. Because of the expense of deploying and operating satellites, satellite-based data is typically gathered by governments or the private sector and tends to be the most consistent in terms of collection protocol and data format. However, there are often limitations in spatial and temporal resolutions, and in availability, as some EO datasets (especially those owned by the private sector) are not freely accessible. At the highest resolution, for example, EO data can distinguish individual trees, but such data is typically not freely available.

Scientific projects and measurement stations are significant sources of agricultural data. Such datasets are highly valuable for their consistency of collection protocol and accuracy of measurements. These datasets may, however, be limited in geographic coverage. The relatively low density of weather stations in Africa, for example, means that African farmers have relatively reduced weather information and ability to predict weather events.[8] Nevertheless, scientific data, particularly pertaining to trees and forests, has immense potential for helping to understand and address global environmental challenges. Data from the dendrochronology lab at CIFOR-ICRAF, for example, are tremendously rich in information, and are useful in modelling various phenomena, from climate change to the hydrological cycle.[9] Information locked in tree rings offers many insights on the right tree for the right purpose in the right location, basic elements critical to successful complex agroforestry systems that support the ability of trees to sequester carbon, store water and nutrients from soil, increase biodiversity and enrich soil organic matter and carbon.

Mobile app-based data collection is growing in popularity and has been deployed throughout the Global South in a variety of contexts. Activating a network of citizen scientists can provide large volumes of data, and while such data may be of variable quality and inconsistent frequency, due to the large numbers of volunteers, such datasets can be very useful.[10],[11]

Traditional knowledge (TK) and Indigenous knowledge (IK) are sources of agricultural data that present unique challenges. Compared with other agricultural datasets, TK and IK may cover the longest timescales, and may be the most narrowly contextualized, pertaining to very focused geographic areas. These datasets may be oral in whole or in part; converting TK and IK to written datasets poses particularly difficulties, both operational and ethical.

Harmonizing the various sources of agricultural data presents significant challenges, although some efforts have been made. At CIFOR-ICRAF, the Land Degradation Surveillance Framework (LDSF) provides a comprehensive method for assessing soil and land health using field-collected data. The method establishes a biophysical baseline at the landscape level. Using data analytics, spatial assessments and a multi-scale approach, the baseline allows monitoring of land degradation and assessment of rehabilitation efforts over time. Investments in capacity for data acquisition and methods of analysis have complemented the LDSF and have resulted in significant datasets pertaining to soil and land health. Crucially, the framework is openly available, contributing to the development of interoperable and complementary datasets from different organizations in different geographic regions. These could be considered minimal steps towards building and scaling a digital commons which, as we describe in the next section, is critical for addressing planet-wide resilience.

A planetary digital commons framework

Rockström and colleagues introduced the concept of a Planetary Commons to describe the interconnected critical biophysical systems determining and regulating the Earth’s environment.[12] The Planetary Commons is distinguished from the existing global commons, the latter being a political and legal construct resting on assumptions of global environmental stability. The Planetary Commons framework treats Earth’s resilience as a goal rather than an assumption. Reaching that goal requires an understanding of Earth systems including the atmosphere, hydrosphere, lithosphere, biosphere and cryosphere,[13] as well as subsystems that include forests, peatlands and other carbon sinks. The Planetary Commons framework leads to “comprehensive stewardship obligations through Earth system governance aimed at restoring and strengthening planetary resilience and justice.”[14]

The importance of data to the Planetary Commons cannot be understated. There can be no solution developed for a problem that is not understood, and the Planetary Commons is a paradigm designed to contextualize solutions for complex challenges on a global scale. It is a framework that relies on evidence available to all and governed in a way that supports evidence-based science and policymaking.

In the context of the Anthropocene epoch and the Planetary Commons there is a critical need for robust governance frameworks that support addressing the interconnected challenges and functioning of Earth’s systems. We propose that a digital commons serves the data-focused needs of the Planetary Commons. A digital commons, properly conceived and operationalized, supports the comprehensive stewardship and integrated management that is a cornerstone of the Planetary Commons framework. This is a paradigm shift from merely managing resource access to ensuring the protection of vital Earth system functions crucial for planetary resilience and justice.

As various forms of knowledge (including TK and IK) increasingly become digital, and AI becomes ubiquitous, the Planetary Commons framework necessitates the incorporation of a digital commons to effectively support custodianship of these planetary systems. For example, accessible and reliable climate data have been recognized as essential for well-grounded resource allocation and decision making.[15]

Contrary to traditional biophysical systems, which can degrade with overuse, a digital commons has the potential to increase in value and utility through wider sharing and ethical use. Digital platforms can provide broader access to information, foster collaboration, and drive innovation. By developing open and accessible digital environments, we can promote sustainable practices and improve decision making for planetary stewardship.

The need for global data and AI governance

Traditional governance principles, such as those proposed by Elinor Ostrom, have been critiqued for their limitations in addressing system design and inclusivity, particularly for marginalized communities.[16] There is a growing recognition of the need for inclusive codesign processes that incorporate diverse stakeholders into the governance of technological systems. Proposals for design principles that build on Ostrom’s framework are emerging, aiming to adapt these principles to the governance of technological innovations.[17]

Through our scientific activities, CIFOR-ICRAF demonstrates a holistic systems approach that leverages the complexity of the interaction between people and ecological systems to increase climate resilience. The complexity of the data needed to ensure effective integration of trees, crops, livestock or fish in landscapes can be supported by databases like CIFOR-ICRAF’s Regreening Africa app or the Tree Globally Observed Environmental Ranges database (TreeGOER). Ensuring the right insights guide productive climate resilient and nature positive food tree portfolio is key.[18]

Recent wider calls for action emphasize the necessity of ensuring that technological advancements promote both sustainability and equity.[19] Notable efforts, such as the Land Degradation Surveillance Framework described above, demonstrate successful integration of scientific research and community engagement. This framework, utilized in over 20 countries on three continents, provides critical land health data that supports decision-making processes related to land restoration. Public datasets on soil health similarly contribute to our understanding of global land and soil commons. An effective digital commons would incorporate similar principles and methods, but on a wider scale, covering all the systems in the Planetary Commons.

Role in innovation

The availability of free and accessible data through digital commons significantly reduces the need for private or open investment to access such data, thus lowering transaction costs.[20],[21] This reduction in costs is crucial for fostering innovation, as it enables a broader range of entities to engage in the development and implementation of new technologies and solutions.[22] The provision of freely accessible data mitigates the barriers to innovation by offering essential information without the financial burden associated with costly transactions. This accessibility is particularly advantageous for small and medium-sized enterprises (SMEs), startups and development organizations, which often face significant constraints in investing heavily in data acquisition. Leveraging data commons allows these entities to innovate more efficiently and effectively.[23],[24]

The reduction in transaction costs is integral to enhancing economic efficiency and innovation, as these costs represent the expenses associated with utilizing goods or services.[25],[26] By lowering these costs, data commons facilitate a more inclusive environment for innovation, enabling diverse stakeholders to contribute to technological advancements and solutions.

Current funding gaps for the commons

Despite the potential benefits offered by data and digital commons, substantial funding gaps persist. Many governmental and private entities have yet to prioritize investments in the development and maintenance of these commons, resulting in their underutilization.[27] Historically, philanthropic organizations have played a crucial role in supporting the creation and upkeep of digital commons. By funding these initiatives, philanthropies can ensure that valuable data and tools remain accessible, thereby fostering an environment conducive to collaborative innovation.[28]

To address the current funding gaps, it is imperative to foster partnerships among the public, private and non-profit sectors. Such collaborations enable stakeholders to pool resources, share expertise, and develop comprehensive strategies for the effective management and utilization of digital commons. These joint efforts can lead to more sustainable and impactful solutions to global challenges by leveraging the diverse strengths and capabilities of participating organizations. Integrating collaborative approaches into philanthropic efforts can enhance the efficacy of investments and ensure that the benefits of digital transformation are equitably distributed across various societal segments.[29]

4. Cultivating optimism and inspiring action

Although modern environmental crises are complex and originate from activities with long historical precedents, there is reason to be optimistic about the power of collective action properly directed and incentivized. The Montreal Protocol on Substances that Deplete the Ozone Layer, passed in the 1980s, is one example demonstrating that planetwide consensus is both possible and highly effective.[30] Science showed the presence of a thinning ozone layer and identified the risks that it posed to the planet. Science also provided alternatives that eliminated the need to rely on known harmful substances. Together, policymakers and scientists managed to create a framework with incentives and alternatives that changed behaviours toward a more sustainable pathway.

Nearly 40 years after the Montreal Protocol, climate change is every bit as threatening to the planet as the thinning ozone layer once was. Although the science of the problem is clear, currently the only effective solution – reducing greenhouse gases – is neither economically nor politically feasible. The increased discussions around geoengineering solutions demonstrates the urgency of the matter, but there are significant unknowns regarding undesirable side effects for most such solutions.[31] Further complicating the matter, and due to its high electricity demand, artificial intelligence has itself been identified as a potential contributing factor to climate change.[32]

Climate change differs from the thinning ozone layer in that the former is a far more complex problem. There are many known causes ranging from burning fossil fuels to agricultural practices, and the phenomenon has components of a feedback loop. Solutions to the problem, whether they be reductions in greenhouse gas emissions or any of the proposed geoengineering processes, have one requirement in common – the availability of data and robust science-led global governance. Earth observation data, ground truthing data, TK and IK, and other forms of data must be collected and managed in a way that enables scientists to understand the problems and propose viable solutions. Data management must be formulated in a way that allows lawmakers and policymakers to trust the science and impose the framework of incentives that will address the issues. This needs to create a robust foundation for an innovation ecosystem in agroforestry that can be agile and responsive to building the resilience we need amidst the uncertainty of the changing climate. The entire process must be inclusive, collaborative and transparent, aiming to benefit the most vulnerable.

The rapid development of artificial intelligence, its reliance on data, and its likely influence on any path towards sustainability, underscores the need for urgency in developing a digital commons that serves the needs of everyone. That can only happen when all stakeholders – including the scientists, policymakers, innovators and funders – prioritize creating and contributing to a digital commons that supports the sustainability of the Planetary Commons. Such an effective planetary digital commons will envision a future where collective intelligence, shared responsibility and inclusive governance drive sustainable outcomes.

[1] Throughout this paper, the term “digital commons” is used broadly to encompass data as well as other forms of knowledge, sociocultural capital and other forms of capital. A subset of a digital commons is a “data commons”, which comprises only data.

[2] Agroforestry is an agroecological approach that involves farmers, livestock, trees and forests at multiple scales – including trees on farms, farming in forests and at forest margins and tree-crop production. See, for example, Nair PKR, 1993. An introduction to agroforestry. Springer Netherlands, Dordrecht.

[3] Crutzen PI and Stoermer EF. 2000. The “Anthropocene”, IGBP Newsletter 41: 12.

[4] Ruddiman W. 2003. The anthropogenic greenhouse era began thousands of years ago. Climatic Change 61: 261–293.

[5] Mahaswa R, Widhianto A, Nurul H. 2021. Eco-agriculture and farming in the Anthropocene epoch: A philosophical review. International Conference on Bio-Energy and Environmentally Sustainable Agriculture Technologies (p. 226). Malang: E3S Web of Conferences.

[6] Ubalijoro E. 2021. A sustainable and equitable digital revolution: Eliane Ubalijoro. One Earth, 4(6), 801-804.

[7] Metz, C., Kang, C., Frenkel, S., Thompson, S.A., & Grant, N. (2024). How tech giants cut corners to harvest data for ai. The New York Times.

[8] Dinku T. 2019. Challenges with availability and quality of climate data in Africa. In Extreme hydrology and climate variability (pp. 71–80). Elsevier.

[9] Gebrekirstos A, Bräuning A, Sass-Klassen U, Mbow C. 2014. Opportunities and applications of dendrochronology in Africa. Current Opinion in Environmental Sustainability 6: 48–53.

[10] Feldmeyer D, Meisch C, Sauter H, & Birkmann J 2020. Using OpenStreetMap data and machine learning to generate socio-economic indicators. ISPRS International Journal of Geo-Information, 9(9), 498.

[11] Li H, Zech J, Hong D, Ghamisi P, Schultz M, & Zipf A, 2022. Leveraging OpenStreetMap and multimodal remote sensing data with joint deep learning for wastewater treatment plants detection. International Journal of Applied Earth Observation and Geoinformation, 110, 102804.

[12] Rockström J, Kotzé L, Milutinović S, Biermann F, Brovkin V, Donges J, Ebbessen J, French D, Gupta J, Kim R, et al. 2024. The Planetary Commons: A new paradigm for safeguarding Earth-regulating systems in the Anthropocene. Proceedings of the National Academy of Sciences 121(5): e2301531121.

[13] Id. at page 5.

[14] Id. (Abstract).

[15] United Nations Environment Programme (2024). The Climate Data Challenge: The Critical Role of Open-Source and Neutral Data Platforms, Geneva.

[16] Robert G, Williams O, Lindenfalk B, Mendel P, Davis LM, Turner S, Farmer C, Branch C. 2021. Applying Elinor Ostrom’s Design Principles to Guide Co-Design in (care) Improvement: A Case Study with Citizens Returning to the Community from Jail in Los Angeles County. International Journal of Integrated Care 21(1): 7. https://doi.org/10.5334/ijic.5569

[17] Stern PC. 2011. Design principles for global commons: Natural resources and emerging technologies. International Journal of the Commons 5(2): 213–232. https://doi.org/10.18352/ijc.305

[18] van Zonneveld M, Kindt R, McMullin S, Achigan-Dako EG, N’Danikou S, Hsieh WH, ... & Dawson IK, 2023. Forgotten food crops in sub-Saharan Africa for healthy diets in a changing climate. Proceedings of the National Academy of Sciences, 120(14), e2205794120.

[19] See FN4.

[20] Nugraha A, Yuniarto A, & Fiedler T 2021. The Role of Data Commons in Reducing Transaction Costs for Innovation. Journal of Innovation Economics, 15(2), 123-145.

[21] Yuniarto A, Nugraha A, & Fiedler T 2021. Accessibility of Data Commons and Its Impact on SMEs. Small Business Economics, 38(3), 567-589.

[22] Fiedler T, Nugraha A, & Yuniarto A 2021. Innovation Through Data Commons: A Comprehensive Review. Technology and Innovation Management Review, 12(4), 234-256.

[23] Garigliotti G, Lyu H, & McIntosh J, 2023. Data Commons and Innovation in SMEs: A Case Study Approach. International Journal of Small Business Management, 50(3), 345-367.

[24] McIntosh J, Edelmann N, & Virkar S, 2023. Leveraging Data Commons for Startups: Opportunities and Challenges. Journal of Business Research, 45(1), 78-99.

[25] Edelmann N, & Virkar S, 2023. Reducing Transaction Costs in Innovation: The Role of Data Commons. European Journal of Innovation Management, 29(1), 45-67.

[26] Lyu H, Garigliotti G, & McIntosh J, 2023. Enhancing Economic Efficiency Through Data Commons. Journal of Economic Perspectives, 37(2), 112-134.

[27] Stanciu V, Putra A, & Yeniasır, M, 2021. Funding Gaps in Digital Commons: An Analysis. Public Administration Review, 81(1), 1-8.

[28] Putra A, Stanciu V, & Yeniasır M, 2022. The Role of Philanthropy in Supporting Digital Commons. Nonprofit and Voluntary Sector Quarterly, 51(2), 234-256.

[29] Sustainability in the Digital Age, Future Earth, and ClimateWorks Foundation. 2022. Dynamic Philanthropy – A Framework for Supporting Transformative Climate Governance in the Digital Age. https://doi.org/10.5281/zenodo.5764443

[30] Gonzalez M, Taddonio KN, Sherman NJ. 2015. The Montreal Protocol: How today’s successes offer a pathway to the future. Journal of Environmental Studies and Sciences 5: 122–129.

[31] Sovacool BK. 2021. Reckless or righteous? Reviewing the sociotechnical benefits and risks of climate change geoengineering. Energy Strategy Reviews 35: 100656.

[32] Kaack LH, Donti PL, Strubell E, Kamiya G, Creutzig F, & Rolnick D 2022. Aligning artificial intelligence with climate change mitigation. Nature Climate Change, 12(6), 518-527.