Charlotte Gohr

Artificial intelligence (AI) holds significant potential to advance Sustainable Development Goals by enabling data-driven insights and optimizations. In this analysis, we review 792 articles that explore AI applications in Sustainable Development Goal-related research. The literature is organized along two dimensions: (1) the disciplinary spectrum, from natural sciences to the humanities, and (2) the focus, distinguishing economic from socioecological content. Deep learning and supervised machine learning were the most prominently applied algorithms for forecasting and system optimization. However, we identify a critical gap: only a few studies combine advanced AI applications with deep sustainability expertise. Sustainability needs to strike a balance between contextualization and generalizability to provide tangible knowledge that will lead to responsible change. AI must play a central role in this process. While expectations for AI’s transformative role in sustainable development are high, its full potential remains to be realized.

UNESCO biosphere reserves serve as learning areas for sustainable development, where preserving ecosystem functionality is an imperative. However, this critical assumption has yet to be thoroughly examined. To address this knowledge gap, we investigated differences in satellite-derived proxies of ecosystem functions in forests between inside and surrounding areas of biosphere reserves, globally. Our findings based on linear mixed effect models show that (i) only 18 of 119 biosphere reserves exhibited higher values for all forest ecosystem function proxies inside the reserves compared to outside, (ii) smaller reserves in fragmented forestscapes were more affected by hot day temperatures, and (iii) greater forest cover correlated with increased ecosystem functioning across all biomes. This study underscores the potential significance of biosphere reserves to biodiversity conservation efforts and the need for the integration of satellite-based, outcome-oriented proxies of ecosystem functions in assessments of protected area effectiveness.

Effective protected areas reflect socio‐ecological values, such as biodiversity and habitat maintenance, as well as human well‐being. These values, which safeguard ecosystem services in protected areas, are treated as models for the sustainable preservation and use of resources. While there is much research on the effectiveness of protected areas in a variety of disciplines, the question is whether there is a common framework that uses remote sensing methods. We conducted a qualitative and a quantitative analysis of 44 peer‐reviewed scientific papers utilizing remote sensing data in order to examine the effectiveness of protected areas. Very few studies to date have a wide or even a global geographical focus; instead, most quantify the effectiveness of protected areas by focusing on local‐scale case studies and single indicators such as forest cover change. Methods that help integrate spatial selection approaches, to compare a protected area's characteristics with its surroundings, are increasingly being used. Based on this review, we argue for a multi‐indicator‐based framework on protected area effectiveness, including the development of a consistent set of socio‐ecological indicators for a global analysis. In turn, this will allow for globally applicable use, including a concrete evaluation that considers the diversity of regional parameters, biome‐specific variables, and political frameworks. Ideally, such a framework will enhance the monitoring and evaluation of global strategies and conventions.