Month: August 2024
Artificial intelligence and its ‘slow violence’ to human rights
Abstract
Human rights concerns in relation to the impacts brought forth by artificial intelligence (‘AI’) have revolved around examining how it affects specific rights, such as the right to privacy, non-discrimination and freedom of expression. However, this article argues that the effects go deeper, potentially challenging the foundational assumptions of key concepts and normative justifications of the human rights framework. To unpack this, the article applies the lens of ‘slow violence’, a term borrowed from environmental justice literature, to frame the grinding, gradual, attritional harms of AI towards the human rights framework.
The article examines the slow violence of AI towards human rights at three different levels. First, the individual as the subject of interest and protection within the human rights framework, is increasingly unable to understand nor seek accountability for harms arising from the deployment of AI systems. This undermines the key premise of the framework which was meant to empower the individual in addressing large power disparities and calling for accountability towards such abuse of power. Secondly, the ‘slow violence’ of AI is also seen through the unravelling of the normative justifications of discrete rights such as the right to privacy, freedom of expression and freedom of thought, upending the reasons and assumptions in which those rights were formulated and formalised in the first place. Finally, the article examines how even the wide interpretations towards the normative foundation of human rights, namely human dignity, is unable to address putative new challenges AI poses towards the concept. It then considers and offers the outline to critical perspectives that can inform a new model of human rights accountability in the age of AI.
Estimation of 100 m root zone soil moisture by downscaling 1 km soil water index with machine learning and multiple geodata
Abstract
Root zone soil moisture (RZSM) is crucial for agricultural water management and land surface processes. The 1 km soil water index (SWI) dataset from Copernicus Global Land services, with eight fixed characteristic time lengths (T), requires root zone depth optimization (Topt) and is limited in use due to its low spatial resolution. To estimate RZSM at 100-m resolution, we integrate the depth specificity of SWI and employed random forest (RF) downscaling. Topographic synthetic aperture radar (SAR) and optical datasets were utilized to develop three RF models (RF1: SAR, RF2: optical, RF3: SAR + optical). At the DEMMIN experimental site in northeastern Germany, Topt (in days) varies from 20 to 60 for depths of 10 to 30 cm, increasing to 100 for 40–60 cm. RF3 outperformed other models with 1 km test data. Following residual correction, all high-resolution predictions exhibited strong spatial accuracy (R ≥ 0.94). Both products (1 km and 100 m) agreed well with observed RZSM during summer but overestimated in winter. Mean R between observed RZSM and 1 km (100 m; RF1, RF2, and RF3) SWI ranges from 0.74 (0.67, 0.76, and 0.68) to 0.90 (0.88, 0.81, and 0.82), with the lowest and highest R achieved at 10 cm and 30 cm depths, respectively. The average RMSE using 1 km (100 m; RF1, RF2, and RF3) SWI increased from 2.20 Vol.% (2.28, 2.28, and 2.35) at 30 cm to 3.40 Vol.% (3.50, 3.70, and 3.60) at 60 cm. These negligible accuracy differences underpin the potential of the proposed method to estimate RZSM for precise local applications, e.g., irrigation management.
Predicting daily maximum temperature over Andhra Pradesh using machine learning techniques
Abstract
Surface Air Temperature (SAT) predictions, typically generated by Global Climate Models (GCMs), carry uncertainties, particularly across different greenhouse gas emission scenarios. Machine Learning (ML) techniques can be employed to forecast long-term temperature variations, although this is a challenging endeavour with few drawbacks, such as the influence of scenarios involving greenhouse gas emissions. Therefore, the present study utilized multiple ML approaches such as Artificial Neural Networks (ANN), multiple linear regression, support vector machine and random forest, along with various daily predicted results of GCMs from Coupled Model Intercomparison Project Phase 6 as predictors and the “India Meteorological Department’s” Maximum SAT (MSAT) as the predictand, to predict daily MSAT in the months of March, April and May (MAM) over Andhra Pradesh (AP) for the period 1981–2022. The results show that ANN outperforms other ML techniques in predicting daily MSAT, with a root mean square error of 1.41, an index of agreement of 0.89 and a correlation coefficient of 0.81. The spatial distribution of hot and heat wave days indicates that the Multiple Model Mean (MMM) underestimates these occurrences, with a minimum bias of 9 and 6 days, respectively. In contrast, the ANN model exhibits much smaller biases, with a maximum underestimation of 3 hot and 2 heat wave days. These findings demonstrate that MMM does not capture the maximum temperatures well, resulting in poor predictability. Further, future temperature projections were analysed from 2023 to 2050, which display a gradual increase in mean MSAT during MAM over AP. This research demonstrates the potential of ML techniques to enhance temperature forecasting accuracy, offering valuable insights for climate modeling and adaptation. The results are crucial for stakeholders in agriculture, health, energy, water resources, socio-economic planning, and urban development, aiding in informed decision-making and improving resilience to climate change impacts.
Geostatistical Kriging Interpolation for Spatial Enhancement of MODIS Land Surface Temperature Imagery
Abstract
Thermal images play a crucial role in various applications, such as environmental monitoring, energy efficiency, and food safety. However, thermal images are often affected by low spatial resolution, limited accuracy, and noise, which reduce their usefulness and effectiveness. This research paper presents a novel approach for enhancing thermal images and optimizing using Kriging Interpolation KI. The proposed KI method combines a metaheuristic optimization algorithm, Particle Swarm Optimization (PSO), with Kriging, a geostatistical method for interpolation and prediction of spatially continuous variables. The proposed KI method has been evaluated on a set of low-resolution Land surface temperature (LST) images of Moderate Resolution Imaging Spectroradiometer (MODIS) satellite and validated with higher resolution LandSat-8 LST. The use of PSO in combination with Kriging provides a powerful tool for efficient and accurate spatial enhancement of thermal images, allowing for the preservation of important thermal features and details while improving the overall quality of the images. The proposed KI algorithm demonstrated the effectiveness of the approach in enhancing the spatial resolution and accuracy of the MODIS thermal images. The results show that the proposed method outperforms traditional statistical LST image enhancement methods, such as DisTrad, TsHarp, and Regression Tree in terms of spatial resolution and accuracy. The proposed method has potential applications in agricultural, metrological, and environmental applications, where thermal images are used to continuously monitor and control temperature-sensitive data.
Building integrated plant health surveillance: a proactive research agenda for anticipating and mitigating disease and pest emergence
Abstract
In an era marked by rapid global changes, the reinforcement and modernization of plant health surveillance systems have become imperative. Sixty-five scientists present here a research agenda for an enhanced and modernized plant health surveillance to anticipate and mitigate disease and pest emergence. Our approach integrates a wide range of scientific fields (from life, social, physical and engineering sciences) and identifies the key knowledge gaps, focusing on anticipation, risk assessment, early detection, and multi-actor collaboration. The research directions we propose are organized around four complementary thematic axes. The first axis is the anticipation of pest emergence, encompassing innovative forecasting, adaptive potential, and the effects of climatic and cropping system changes. The second axis addresses the use of versatile broad-spectrum surveillance tools, including molecular or imaging diagnostics supported by artificial intelligence, and monitoring generic matrices such as air and water. The third axis focuses on surveillance of known pests from new perspectives, i.e., using novel approaches to detect known species but also anticipating and detecting, within a species, the populations or genotypes that pose a higher risk. The fourth axis advocates the management of plant health as a commons through the establishment of multi-actor and cooperative surveillance systems for long-term data-driven alert systems and information dissemination. We stress the importance of integrating data and information from multiple sources through open science databases and metadata, alongside developing methods for interpolating and extrapolating incomplete data. Finally, we advocate an Integrated Health Surveillance approach in the One Health context, favoring tailored and versatile solutions to plant health problems and recognizing the interconnected risks to the health of plants, humans, animals and the environment, including food insecurity, pesticide residues, environmental pollution and alterations of ecosystem services.
The AIR and Apt-AIR Frameworks of Epistemic Performance and Growth: Reflections on Educational Theory Development
Abstract
The nurturing of learners’ ways of knowing is vital for supporting their intellectual growth and their participation in democratic knowledge societies. This paper traces the development of two interrelated theoretical frameworks that describe the nature of learners’ epistemic thinking and performance and how education can support epistemic growth: the AIR and Apt-AIR frameworks. After briefly reviewing these frameworks, we discuss seven reflections on educational theory development that stem from our experiences working on the frameworks. First, we describe how our frameworks were motivated by the goal of addressing meaningful educational challenges. Subsequently, we explain why and how we infused philosophical insights into our frameworks, and we also discuss the steps we took to increase the coherence of the frameworks with ideas from other educational psychology theories. Next, we reflect on the important role of the design of instruction and learning environments in testing and elaborating the frameworks. Equally important, we describe how our frameworks have been supported by empirical evidence and have provided an organizing structure for understanding epistemic performance exhibited in studies across diverse contexts. Finally, we discuss how the development of the frameworks has been spurred by dialogue within the research community and by the need to address emerging and pressing real-world challenges. To conclude, we highlight several important directions for future research. A common thread running through our work is the commitment to creating robust and dynamic theoretical frameworks that support the growth of learners’ epistemic performance in diverse educational contexts.
Wave climate around New Caledonia
Abstract
Pacific islands are widely exposed to several strong wave events all year long. However, comprehensive analyses of coastal vulnerabilities to wave climates and their extremes are often lacking in those islands. In the present paper, the wave climate around the reef of New Caledonia is analyzed using a 28-year simulation performed with the Wave Watch III model, and accounting for realistic wind intensity forcing from tropical cyclones. Four mean wave regimes are defined with clustering methods, and are shown to vary along the reef depending on its main orientation. The western reef is mostly exposed to energetic south-western swells (significant height over 1.5 m, peak period of ~ 12 s) generated in the Tasman Sea that are reinforced during austral winter. The northern sector and the Loyalty Islands, are hit by shorter waves (~ 8 to 9 s period) coming from the south-east to the north-east, with height ranging on average from 0.8 m in the Loyalty Channel to 1.5 m at the northern tip of the Grande Terre reef. These waves mainly result from the south-eastern trade winds that blow over the central south-western Pacific all year long. In austral summer, additional swell remotely generated by both the extra-tropical westerlies and the north-eastern trade winds of the northern hemisphere reach the north-eastern reef of the archipelago. These wave regimes also strongly vary in response to the interannual El Niño-Southern Oscillation. El Niño events tend to increase the frequency of the south-western swell regime in austral spring and fall, and of the south-eastern trade wind waves in austral summer. In contrast, during La Niña, waves generated in the northern hemisphere are more likely to reach New Caledonia all year long. Finally, extreme wave events and their return periods were assessed. Wave amplitude reaching 7 m is estimated to occur every 100 years. On the west and southern tip of the Grande Terre reef, extreme waves are 80% of the time westerly waves generated by storms in the Tasman Sea or in the Coral Sea, while on the eastern reefs (Loyalty Islands and Channel), 70% of the extreme wave episodes are associated to tropical cyclone-induced waves. During La Niña episodes, more tropical cyclones pass by New Caledonia, increasing their contribution to extreme wave events along the western and southern coasts of the island. Conversely, in El Niño conditions, the exposure to tropical cyclone-induced waves is predominantly concentrated on the northeastern side.
Hierarchical machine learning models can identify stimuli of climate change misinformation on social media
Abstract
Misinformation about climate change poses a substantial threat to societal well-being, prompting the urgent need for effective mitigation strategies. However, the rapid proliferation of online misinformation on social media platforms outpaces the ability of fact-checkers to debunk false claims. Automated detection of climate change misinformation offers a promising solution. In this study, we address this gap by developing a two-step hierarchical model. The Augmented Computer Assisted Recognition of Denial and Skepticism (CARDS) model is specifically designed for categorising climate claims on Twitter. Furthermore, we apply the Augmented CARDS model to five million climate-themed tweets over a six-month period in 2022. We find that over half of contrarian climate claims on Twitter involve attacks on climate actors. Spikes in climate contrarianism coincide with one of four stimuli: political events, natural events, contrarian influencers, or convinced influencers. Implications for automated responses to climate misinformation are discussed.
A life engineering perspective on algorithms, AI, social media, and quantitative metrics
Abstract
This academic paper delves into the captivating intersection of life engineering and algorithms, artificial intelligence (AI), social media, and quantitative metrics on human life, through a comprehensive review of three thought-provoking books. In each critical review, the authors add their own thoughts and impressions, as Computer Science graduates and scholars, illustrating the impact that these eye-opening books have on them. The first book, “Weapons of Math Destruction” by Cathy O’Neil, delves into the hidden dangers of algorithmic decision-making. O’Neil uncovers how algorithms can perpetuate discrimination, biases, and unfairness in domains such as education, advertising, criminal justice, employment, and finance, and emphasizes the need for ethical considerations, transparency, and human judgment in algorithmic systems. The second book, “Atlas of AI” by Kate Crawford, takes a multidimensional approach to AI beyond mere algorithms and deep learning. Crawford addresses issues such as labor exploitation, surveillance technologies, classification systems, wealth concentration, and environmental consequences due to AI. The book calls for responsible and ethical considerations in the development and usage of AI. Shoshana Zuboff’s “The Age of Surveillance Capitalism” is the third book, focusing on the pervasive influence of tech giants like Google and Facebook. Zuboff exposes the dynamics of surveillance capitalism, wherein personal data is extracted and exploited for economic gains. The book illuminates how this form of capitalism erodes privacy, reshapes societal structures, and challenges democratic norms. Illustrating the essence of these disruptive narratives and the tense dialogue taking place between ethicians or scholars and technology developers, this research examines the profound social, economic, and environmental implications brought forth by these transformative technologies. Ultimately, the paper advocates for the embrace of responsible and ethical technology development that not only safeguards the well-being of individuals but also fosters a harmonious coexistence between humans and machines amidst the winds of disruption.