Contextualizing the Research Agenda for AI and Sustainable Development

In 2015, the Sustainable Development Goals (SDGs), representing 17 interconnected themes across five pillars (planet, people, peace, prosperity, and partnerships), were adopted by UN member states with an aim to achieve global sustainable development by 2030 (Biermann et al., 2017). With less than 5 years remaining to achieve them, progress is well off-target, and in some instances, gains that were made have been reversed as a result of increasing global complexity and uncertainty, conflict, economic and social instability, as well as widespread public health crises (ITU, UNDP, 2023; Sachs et al., 2019).

It is widely recognized that deep and urgent transformations are required to achieve a sustainable future, and a critical way to do so is by leveraging digital technologies to accelerate development (Gouvea et al., 2018; ITU, UNDP, 2023; Sachs et al., 2019; Seele & Lock, 2017). The advent of new technologies, systems and devices, the explosion of computational power, and the development of innovative data management and processing systems, offers new opportunities to achieve sustainability while mitigating risks and overcoming barriers. As suggested by Nambisan et al. (2017), digital technologies are fundamentally a resource to achieve our goals; therefore, a key issue is how to exploit these effectively to amplify efforts toward sustainable development.

The role of artificial intelligence (AI), in particular, is commanding significant attention across global scholarly, policy, and public discourse. As the world navigates the twin transitions of sustainable development and digital transformation, AI is increasingly touted as the cornerstone of innovation and a key enabler to bridge these domains. In a recent review by Gohr et al. (2025), 792 publications linking AI to the SDGs were examined, finding that AI is often used as a tool for forecasting, system optimization, data mining, and accelerated experimentation. However, the authors note that most research concerns the technical aspects of AI or addresses specific SDGs, and the use or impacts of AI on “complex sustainability challenges” are lacking.

Furthermore, even though AI offers transformative potential across all domains of human life (e.g., healthcare, education, and environment), there are major questions and concerns about its ethical implications, societal impacts, and risks, including the capacity to disrupt labor markets, amplify biases, and reduce human autonomy (Khogali & Mekid, 2023). The discourse surrounding AI is marked by duality; it is highlighted as a solution to global challenges, yet also understood to pose novel risks, especially for those who are already marginalized, vulnerable, or embedded in contexts of low development. To overcome these issues, it is suggested that a global view of AI ethics be taken that includes an understanding of the distinct regional and national impacts of AI (i.e., it is culturally bound) and that it acknowledged that AI can exacerbate inequality (as well as potentially decrease it; Hagerty & Rubinov, 2019). As Tripathi et al. (2024) argue, addressing sustainability challenges through AI requires hybrid approaches that integrate data-driven methods with expert knowledge, inclusive governance, and ethical safeguards.

This special issue sought submissions that explored the role of AI in global sustainable development by explicating and examining its potential to drive progress while also sensitizing to local, ethical, and social challenges. The 10 included submissions (three research articles, three research notes, and four colloquium) all share key features; a critical approach to AI ethics with a focus on data colonialism, justice, and technosolutionism, an examination of the societal implications, ethical governance practices and participatory approaches to AI, and the inclusion of case studies outlining sensitive and novel applications of AI for the SDGs, particularly in the Global South. Two key contributions identified across the special issue are (1) Power Dynamics and AI, and (2) The Role of Human Agency and Inclusion, both of which are subsequently outlined and the submissions introduced.

Power Dynamics and AI

The first key contribution of this special issue is a critique of the power dynamics embedded in AI that are often (intentionally or not) invisible and the integration of postcolonial perspectives with discourses of modernization or technosolutionism. In Siad and Sagar's (2025) paper, the dual opportunities and risks of AI are examined through the case of an industry-led digital identity system in Africa. The research illustrates that even solutions argued as “tech for good” can intensify existing inequalities through extractive methods, particularly in contexts where there are regulatory gaps. Racine (2025) highlights AI's potential for good within Global Majority healthcare settings while warning that its rapid deployment raises ethical concerns regarding perpetuating harm, entrenching biases, and widening global divides. Racine introduces and advocates for Reparative Algorithmic Impact Assessments (R-AIAs) that are grounded in decolonial and intersectional values to promote data inclusivity and sovereignty, foster participatory governance, and redress systemic biases, inequities, and power imbalances. De Matas (2025) similarly focuses on healthcare settings, conceptualizing AI systems as “narrative machines” that are embedded with the worldviews and values of developers (who are generally located in the Global North), which leads to bias and injustice by excluding local knowledge systems and epistemologies. Ooko (2025), focusing on AI in African public service delivery, also highlights how AI is intertwined with Western ideologies and argues that a lack of thoughtful adoption can lead to data coloniality, algorithmic control, and exclusion. This is reinforced in the manuscript of Dwivedi (2025), where the supposed neutrality of AI is called into question, and it is suggested that, in and of themselves, AI tools cannot overcome social hierarchies and achieve equality as they are encoded with pre-existing biases.

The Role of Human Agency and Inclusion

Another main theme of the special issue is the importance of human participation in AI design and deployment. Although Dwivedi (2025) argues that AI can reify existing unequal systems, they also suggest that this can be minimized by centralizing Human-in-the-Loop (HITL) frameworks. Liu et al. (2025) also emphasize the importance of HITL frameworks, particularly applying this to the case of enhanced social simulations (Agent-Based Modeling [ABM] with Large Language Model [LLM]-driven autonomous agents). They argue that to overcome the technical and ethical challenges of AI systems, diverse, and marginalized voices must be included to allow AI tools more nuanced, human-like decision making. Du et al. (2025) take a human-centred approach in their research on AI driven VR tools for learning and immersion in cultural heritage, highlighting the critical role of user experiences and inclusion in the development of innovative technologies for the SDGs. Alongside the potential for better decision making and more inclusive user experiences, Carpentier (2025) argues that people (rather than technology) have to be centralized because knowledge and wisdom can be overshadowed by the dominance and prioritization of data, resulting in the diminishment of human agency. Odermatt et al. (2025) also examine the impact of a data central world, examining the impacts of automatically collecting, storing, and analyzing users’ data traces. Notably, a sense of “dataveillance” can lead to withdrawal from digital communication behavior. Relatedly, Huang and Wu (2025) find that when AI services fail, the public does not treat the systems as responsible agents but rather expects that corporations take accountability. Therefore, industry culpability, including human oversight, ethical frameworks, and responsible governance, is seen by the public as paramount to mitigate the harms of technology.

Conclusion

This special issue underscores the many possibilities of AI in advancing sustainable development, while critically interrogating its ethical, social, and cultural implications. The 10 manuscripts in this issue collectively advocate for a shift from technosolutionist narratives toward inclusive, participatory, and context-sensitive approaches that foreground human agency, inclusion, and ethical governance. By centering voices from the Global South (e.g., De Matas, 2025; Ooko, 2025; Racine, 2025; Siad & Sagar, 2025) and highlighting key frameworks such as the Reparative Algorithmic Impact Assessments (Racine, 2025), the special issue offers a timely and necessary reorientation of the AI for SDG research agenda, one that is attuned to complexity, equity, and justice. By critically interrogating power dynamics, addressing structural inequities, and supporting participatory AI design and implementation, this diverse set of colloquia, research notes, and research articles each highlight distinct ways in which we can ensure that AI serves as a catalyst for sustainable development rather than a driver of new forms of exclusion and harm.