The building and construction sector accounts for around 39% of global carbon dioxide emissions and remains a hard-to-abate sector. We use a data-driven analysis of global high-level climate action on emissions reduction in the building sector using 256,717 English-language tweets across a 13-year time frame (2009–2021). Using natural language processing and network analysis, we show that public sentiments and emotions on social media are reactive to these climate policy actions. Between 2009–2012, discussions around green building-led emission reduction efforts were highly influential in shaping the online public perceptions of climate action. From 2013 to 2016, communication around low-carbon construction and energy efficiency significantly influenced the online narrative. More significant interactions on net-zero transition, climate tech, circular economy, mass timber housing and climate justice in 2017–2021 shaped the online climate action discourse. We find positive sentiments are more prominent and recurrent and comprise a larger share of the social media conversation. However, we also see a rise in negative sentiment by 30–40% following popular policy events like the IPCC report launches, the Paris Agreement and the EU Green Deal. With greater online engagement and information diffusion, social and environmental justice topics emerge in the online discourse. Continuing such shifts in online climate discourse is pivotal to a more just and people-centric transition in such hard-to-decarbonise sectors.

Identifying drivers of climate misinformation on social media is crucial to climate action. Misinformation comes in various forms; however, subtler strategies, such as emphasizing favorable interpretations of events or data or reframing conversations to fit preferred narratives, have received little attention. This data-driven paper examines online climate and sustainability communication behavior over 7 years (2014–2021) across three influential stakeholder groups consisting of eight fossil fuel firms (industry), 14 non-governmental organizations (NGOs), and eight inter-governmental organizations (IGOs). We examine historical Twitter interaction data (n = 668,826) using machine learning-driven joint-sentiment topic modeling and vector autoregression to measure online interactions and influences amongst these groups. We report three key findings. First, we find that the stakeholders in our sample are responsive to one another online, especially over topics in their respective areas of domain expertise. Second, the industry is more likely to respond to IGOs’ and NGOs’ online messaging changes, especially regarding environmental justice and climate action topics. The fossil fuel industry is more likely to discuss public relations, advertising, and corporate sustainability topics. Third, we find that climate change-driven extreme weather events and stock market performance do not significantly affect the patterns of communication among these firms and organizations. In conclusion, we provide a data-driven foundation for understanding the influence of powerful stakeholder groups on shaping the online climate and sustainability information ecosystem around climate change.

The ongoing global race for bigger and better artificial intelligence (AI) systems is expected to have a profound societal and environmental impact by altering job markets, disrupting business models, and enabling new governance and societal welfare structures that can affect global consensus for climate action pathways. However, the current AI systems are trained on biased datasets that could destabilize political agencies impacting climate change mitigation and adaptation decisions and compromise social stability, potentially leading to societal tipping events. Thus, the appropriate design of a less biased AI system that reflects both direct and indirect effects on societies and planetary challenges is a question of paramount importance. In this paper, we tackle the question of data-centric knowledge generation for climate action in ways that minimize biased AI. We argue for the need to co-align a less biased AI with an epistemic web on planetary health challenges for more trustworthy decision-making. A human-in-the-loop AI can be designed to align with three goals. First, it can contribute to a planetary epistemic web that supports climate action. Second, it can directly enable mitigation and adaptation interventions through knowledge of social tipping elements. Finally, it can reduce the data injustices associated with AI pretraining datasets.

Energy research seeking to influence policy in low- and -middle-income countries (LMICs) is often funded by – and conceptualised by authors in - institutions from high-income countries (HICs). Research agendas and policy recommendations determined in HICs potentially yield the most influence on policymaking for LMICs. This leaves a multidimensional gap in how LMICs frame, contextualise, evidence and enact policy processes. The unique contribution of this paper is analysing the dynamics of prevalent energy research on LMICs through a multi-method approach using bibliometric, network science and regression-based techniques. An innovative data-driven framework was established using a sample of 6,636 papers from the Web of Science database, combined with journal impact data from Scimago Journal Ranking and country economic data from the World Bank. Results show the existence of a cycle of imbalances across research practices. Most papers recommending energy policy for LMICs have a first author based in a HIC, funded by a HIC institution. Total citations of articles on energy policy in LMICs increase with the GDP of the first author’s country (a 1% increase in GDP is correlated with a 0.68% increase in total citations). Funders support authors based in countries of the same income band as them, or higher. Therefore, we recommend revising research practices and HIC funding policies to place local actors and knowledge at the heart of energy policy research, enabling high-impact policymaking in LMICs.