Tropical forest ecosystems play a major role in shaping the outcomes of both the climate and biodiversity crises. Understanding how these ecosystems are impacted by both crises in terms of health, biodiversity and forest cover position monitoring as an essential practice that can reveal long-term trends, contribute to new knowledge and inform environmental planning and policy (Mazotti, 2008).
In recent years, remote sensing and machine learning have emerged as invaluable tools supporting the understanding and monitoring of forest ecosystems to aid conservation efforts; these techniques enable the observation of environments and subsequent collection and analyses of their data from a distance.
The vast majority of forest conservation research is centred on Indigenous lands, yet often operates at a significant remove from Indigenous communities themselves, who are the traditional custodians of the conservation priority areas. In the wake of the colonial project, there has been a long-standing acrimonious relationship between the beliefs and practices of traditional communities and those of formal forest science (Parrotta and Trosper 2012) resulting in traditional practices being disrespected and under-promoted in conservation (Cobbinah., et al 2015). Indigenous knowledge is often the subject of ‘scientific study', but seldom shapes, influences or leads scientific processes (Yahaya, 2012).
Although remote sensing and machine learning techniques have provided necessary support for biodiversity monitoring, these methods can widen the disconnect between conservation projects and indigenous communities and weaken important links and connections to essential local knowledge. Remoteness and removedness coalesce into reduction. The use of state-of-the-art data science must not perpetuate colonial patterns of knowledge displacement and marginalisation in the name of scalability, financial gain and power.
Unfortunately, the latter motivations have become key underpinnings (consciously or subconsciously) of work currently at the nexus of technology and conservation. I am an affiliated researcher on the Smart Forests project, led by my Co-Supervisor Jennifer Gabrys. The project aims to investigate the "social-political impacts of digital technologies that monitor and govern forest environments". In the accompanying publication 'Smart forests and data practices:From the Internet of Trees to planetary governance', Gabrys asks how the "digitalisation [of forests] not only change the understandings of environments but also generate different practices and ontologies for addressing environmental change". Consideration is also given to "how these digital and data-based reworkings of forests potentially lead to significant changes in environmental engagement and planetary governance. From mass sky-borne tree planting facilitated by drones and unmanned aerial vehicles (UAVs) to machine learning powered forest fire alert systems, forests are truly becoming "technologised sites of data production", as introduced by Gabrys.
Counter to the way forests are perceived within the majority of machine learning and conservation projects, forests are not just sites of data, they are homes, sources of spiritual solace and cultural importance as well as sites of a wealth of medicines and resources for local communities. Even for those researchers of liberal disposition, who work to advocate for better livelihoods for forest fringe communities, ideas about what communities need or want are imposed without input from those communities themselves.
We are in need of a paradigm shift. One that ceases the acknowledgement and validation of only science or knowledge creation that occurs within institutional walls but instead has the capacity to see science for what it is, a process of knowledge production (Liebenberg., et al 2021), which transcends the academe.
My PhD research will take an interdisciplinary approach, combining Machine Learning (ML), Bioacoustics, Forest Ecology, Indigenous Knowledge (IK) and Sociology to investigate the role of technology in forest conservation. Influenced not only by the work of Gabrys but also the work of my Supervisor Professor Alan Blackwell, the principles of Radical AI, Critical Theory and Critical Technical practice, my doctoral research, through co-creation and collaboration with Indigenous and local communities, intends to develop and implement a framework that explores new modes of scientific inquiry, challenges currently accepted structures of power and advocates for truly intersectional and interdisciplinary research for real-world impact.
Deep Listening is a practice developed by the late Pauline Oliveros that "explores the difference between the involuntary nature of hearing and the voluntary, selective nature of listening". The intention of the practice is to "cultivate a heightened awareness of the sonic environment, both external and internal, and promotes experimentation, improvisation, collaboration, playfulness, and other creative skills vital to personal and community growth". Although growing from conceptual imaginings, my work seeks to practically embody these themes through action based research, real world application, and collaborative and participatory methods. Acoustic sensing through the deployment of sensors in the canopy of Ghanaian forests will lead to the collection of forest soundscapes that when listened to and analysed by machine learning algorithms, will provide insights on species richness, forest health, the prevalence of culturally important species and patterns of conservation and cultural value given to regions of the forest. Importantly, through interviews, listening exercises and workshops, the design of the deployment and the retrieval of specific insights (the "selective listening") will be co-created and co-lead by the community who know the forest best.
This work is reflexive, not static. Counter to dominant AI discourse, "which makes it exceptionally difficult to conceptualise alternatives to the field's prevailing ideas" (Agre, 1997) constant critique and reflection will be a key part of my methodology. This practice of Deep Listening sits not as a convenient metaphor for this work, but will guide a practice of radical attentiveness, vulnerability and honesty. Forcing me to step back and acknowledging that I don’t have an understanding of all realities and experiences and giving space to those whose voices and narratives are very often silenced or ignored. As Agre mentions, this work "will not always be comfortable, but it will be productive nonetheless, both in the esoteric terms of the technical field itself and the exoteric terms by which we ultimately evaluate a technical field's contribution to society".
Paraphrasing Thomas Sankara, the inclusion and centring of marginalised communities is not one of charity, but rather a necessary basis for the triumph of our actions, policies and solutions in response to the climate emergency. If technology is to be a key and useful facilitator of this triumph, it must give way urgently to praxis based equity, deep unlearning and imagination untethered from the practice of ignorance or arrogance in the name of power.