# Can Artificial Intelligence Counter Carbon Hegemony? **Date de l'événement :** 13/11/2025 * Publié le 13/11/2025 ### Date 01/11/2025 ## Chapô **In the United States, the Yale Center for Environmental Justice, headed by the author of this article, is developing artificial intelligence tools that enable communities most exposed to the impacts of global warming to document the damage they suffer and hold those responsible to account. The goal is to make these communities intellectual, moral and economic leaders in the fight against carbon hegemony.** ## Corps du texte Gramsci’s Prison Notebooks are a touchstone of strategy for how subordinate, oppressed, or powerless groups can create change in the most difficult of circumstances. The Notebooks broke new ground in defining hegemony as a sort of ‘surround sound’ that envelops us all, the collection of attitudes, ideas, and economic, social and physical infrastructures that manufacture and enforce consent while radically constraining political action. In this context, ideology is a set of beliefs that ‘make sense’ – that, authored by dominant groups, combine to reinforce hegemony or are contested by popular movements that emerge to undermine hegemony. In Gramsci’s powerful conception, ‘common sense’ is the key battleground. It’s where ideology runs afoul of the contradictions of the real world – the ways in which dominant ideas, such as ‘fossil fuels are the best source of energy’, have to confront what people see and experience in the real world, namely ‘fossil fuels are causing immense damage to the world’. The many rifts in ‘common sense’ are precisely where counter-hegemony can coalesce, challenging the status quo with a more coherent, conscious and progressive worldview. What AI and climate share ------------------------- There is an immense opportunity today for counter-hegemonic work at the intersection of artificial intelligence (AI) and climate change. Both labelled an ‘existential threat to humanity’, AI and climate change are both composed of extreme contradictions – things we are actively creating that will also annihilate us – problems where common sense (in Gramscian terms) absolutely falls apart. Take, for instance, how Fidji Simo, OpenAI’s CEO of Applications, recently introduced herself to the world with this open letter (written in bold for emphasis): ‘If we get this right, AI can give everyone more power than ever.’ Yet the path she outlines to that power never actually engages with power structures. In the same open letter she says, ‘In the future, people will be able to build new things without waiting for permission, capital, or credentials,’ but the only thing driving that change is the data and logic supplied by AI. In the existential threat of climate change too there is a gap between a mainstream discourse, focused largely on the data and logic of atmospheric science and zero-carbon technologies, and the enormous political, economic and social disruption necessary to actually solve the problem. Informed by diverse communities and Indigenous cultures around the world, the movement for climate justice has engaged this gap and been the leading counter-hegemonic force in the climate space for decades. In partnership with some of these communities, the Yale Center for Environmental Justice has, for the last 18 months, been actively developing AI to support this work, to amplify grass-roots voices in the evolution of ‘common sense’, and to help frontline communities build power. This article discusses the rationale for that work as well as tools that can explicitly support power-building and the erosion of ‘carbon hegemony’ through social, economic and policy mobilisation. It aims to understand why it is still so rare for AI to be used in such contexts, including the paradoxes inherent to counter-hegemonic use of AI, and how working at the intersection of AI and climate change has the potential to exploit those paradoxes to grow consciousness of the structures and ideologies needed to build grass-roots power in this era and shape outcomes that are more just. AI and the futility of ‘intelligence’ in the climate debate ----------------------------------------------------------- In general, the uses of AI fall into three categories: AI can help build understanding by collecting, summarising and analysing information; it can help generate new ideas, and it can help drive action by automating, optimising or speeding real world activities (Klocek, 2025). There is no shortage of applications where AI, indeed, holds great promise within the current ‘scientificized’ ideological frame to better understand climate change, generate new solutions, and act in new ways. These advances are welcome but, well before the emergence of generative AI, there was already no shortage of technological solutions for mitigating the causes of climate change. Solar electricity today is cheaper than that generated by many old coal plants and there are behavioural or technological zero-carbon alternatives for almost every major human activity. The real challenges to mitigating climate change are no longer technological but rather social, economic and political. The entire planet is in the grip of a carbon (or fossil fuel) hegemony, supported by a persistent physical infrastructure (roads, planes, types of food production … the list is long) and an ideological infrastructure that includes global markets, political structures, diverse forms of ‘consent’ and co-optation that wed diverse groups to existing, harmful systems, and even extreme forms of obfuscation. Note, for example, that even after 30 Conferences of the Parties, the text of the United Nations Framework Convention on Climate Change makes no mention of the words ‘fossil fuels’. The greatest challenges are thus not about the science of climate change, a lack of technological innovation, or even a shortage of ways to act. The greatest challenges are, as Al Gore started pointing out almost 30 years ago, a lack of political will and, as any climate policymaker or activist will tell you, a shortage of political, economic and social power to keep fossil fuels in the ground and out of the sky. Climate justice and carbon hegemony ----------------------------------- By almost any definition of justice, those most affected by an injustice should have significant if not dominant voice in the process of remedy. Nowhere is the disjuncture between this fundamental ethic more extreme than in the case of climate change. It is hard to even conceptualise how the victims of recent floods in Texas or the millions exposed to a month of average daytime temperatures above 40 degrees Celsius in India could have meaningful say over policies across the globe that continue to allow dozens of gigatonnes of excess carbon dioxide and other greenhouse gases emitted predominantly from fossil fuels. While prevention in the form of greenhouse gas mitigation is ideal, the voices of frontline climate communities is also often absent in efforts to secure funds for and implementation of adaptation responses to the most severe impacts. In this context, the ethical chasm is part and parcel of a physical, economic and sociocultural hegemony that supports the unceasing increase in emissions. ‘Common sense’ is defined by the energy and agricultural infrastructure of daily life, a chaotic understanding of what drives climate impacts like heat and floods, and more than a half century of active disinformation about the role of fossil fuels in causing climate disasters. The climate justice movement, which actively centres power in the climate change narrative, is the most widespread counter to this hegemony. Its diverse strands share stories that debunk the inevitability of climate change by highlighting winners and losers across and within nations, generations, species (including the human species), and geographies from local to global. The ideological counter at play is simply that there are name-able, responsible parties at multiple scales and that climate action depends on shifting dominant forms of carbon power and ideology in favour of those most at risk from climate change, including nature itself. That dominant ideology is being contested in a major way by having frontline communities play an important intellectual, moral and eventually economic leadership role in the evolution away from fossil fuels. And this position is being established on at least three major ‘terrains’. The first terrain is definition of the climate ‘problem’, or how climate change is described and attributed to causes and responsible par t ies (which remains a significant field of material and ideological debate). Climate justice activists seek to grow the understanding of how much climate change is already underway and the breadth of its impact on human and natural communities. Increased understanding and consciousness is central to exposing the moral and ideological contradictions of the dominant systems. The second terrain is climate mitigation, or what actions will be taken to drop greenhouse gas emission functionally to zero. The movement highlights the centrality of frontline communities to the evolution away from a fossil fuel economy (aka just transitions). Much is made of the market valuations of the fossil fuel sector (USD 7 – 10 trillion) and the enormous, proven reserves of fuel that can never be used if we are to avoid climate catastrophe. But the ‘cashflow’, or day-to-day monetary transactions of supporting industries (from shipping to catering and from truck drivers to energy traders) undoubtedly exceed that market valuation every year and serve as a sort of economic ‘cement’ binding us all to the carbon hegemony. The subsidies to fossil fuel industries alone for example amount to more than USD 7 trillion per year. The third terrain is climate adaptation, or how communities and people are adapting to the changing climate at multiple scales in time and space, but also with respect to our relations with the economic, political and natural world. Willingly or unwillingly, climate change is driving transformation in communities at every scale, and climate justice seeks to grow the power of those under the most intense pressure to adapt, to shape adaptive responses and the future. While frontline communities can stake claim to moral and ethical leadership by dint of their direct experience of the impacts of climate change, there are formidable informational and ideological barriers to their contestation in each of these terrains. And it is precisely these barriers that AI can help climate justice communities breach. From common sense to good sense ------------------------------- Hegemony relies on fragmentation and contradiction to maintain a ‘common sense’ counter to the interests of oppressed groups. In the case of climate change, conceptions of the three terrains described above are too diverse, contradictory and numerous to catalogue here but it is no coincidence that the fossil fuel industry spent at least USD 5.8 billion on climate obfuscation between 2020 and 2022 in the United States alone. Bad information is key to maintaining the status quo in the face of otherwise overwhelming evidence of harm. The result of this flood of falsehood is that most actors in the climate field, including climate justice communities, lack the capacity to document the harms they face and to hold responsible parties accountable. When commitments to predict impact, to mitigate harm or to support adaptation are made, they can falter on shifting narratives of causality and blame. The science of the polar vortex can’t break through in the mass media discourse around ‘How can the planet be warming when we have unusually cold winter days? Is it really true that it takes as much fossil fuel energy to build a solar electric plant than to run a coal-fired generator for thirty years? And why can’t insurance companies insure homes in Florida anymore?’ A fragmented and contradictory ‘common sense’ is thus a robust and abundant component of carbon hegemony. Counternarratives can nevertheless build on the fragmentation and discontinuities within ‘common sense’ to develop a ‘good sense’ that opens new counter-hegemonic opportunities. AI can help frontline communities add an informational and ideological base to the moral stance they already occupy with respect to climate change and build counter-ideologies that place them at the centre of climate debates. In practice, the Yale Center and its partners have so far developed and deployed prototypes of tools for impact and mitigation. The first of these evolved in response to a need for a scalable ‘organiser’s toolkit’ in Connecticut. It provides rudimentary AI to help communities respond formally to otherwise opaque and inaccessible permitting processes. It also provides rudimentary, street-level data on future climate risk. The second tool, Public Participation in Environmental Rule-Making (‘Pepper’), spins up community-specific AI instances across the United States. Organisations, organisers and community leaders can ‘roll their own’ versions that help their members make personalised and legally actionable public comments on new regulatory proposals. A subsidiary version (‘Dr Pepper’) is tuned to generate draft public comments for submission that are based on their own much more detailed research and policy analysis. A third, very early prototype, climate Action & Information for communities (‘cAIc’) is a global version of Pepper, designed to operate in extremely diverse political and policy environments and serve a range of community mobilisation models. AI tools relevant to the climate adaptation column, are a critical next step. Frontline communities need power in the debates over climate impact and mitigation but are also in dire need of capacity for climate adaptation in real time as well as the ability to bring their own wisdom to the fights that will shape their future. The paradox(es) of counterhegemonic AI -------------------------------------- ‘\[Praxis\] must be a criticism of “common sense”, \[...\] It is not a question of introducing from scratch a scientific form of thought into everyone’s individual life, but of renovating and making “critical” an already existing activity’, wrote Gramsci. New understandings in Gramscian terms emerge not from a blank slate, but from the contradictions that already operate in people’s daily experience, and the deployment of counter-hegemonic tools has surfaced at least two types of internal conflicts that communities of practice are already experiencing. The first relates to the specific engagement with climate change and carbon hegemony; the second with the ways in which AI may inherently exacerbate inequality. A critical near-term issue among communities who have started using the ‘climate justice AI’ has been the ways in which environmental justice communities have been among those most affected by the irresponsible deployment of data centres. An egregious example is the new xAI facility in Memphis, TN, which boasts of having built the largest AI computing facility in the world in less than four months. This speed was achieved by using a regulatory loophole and installing ‘temporary’ gas turbines in a predominantly African-American neighbourhood already plagued with bad air quality (Chow, 2025). The environmental problems with AI deployment extend beyond injustice. The demand for energy projected by AI companies has led to claims of ‘game over’ for greenhouse gas emissions control as the imperative of growing AI (and feeding its energy needs) is seen as more critical than the threat of global warming. Such extravagant energy claims are belied by the fact that data centres overall (including AI) are expected to contribute only about 8 per cent of growth in electricity demand over the next five years. There are legitimate capacity issues related to AI deployment but the frenzy and implied trade-offs with justice and climate mitigation are in the well-worn tradition of strategic over-statement by polluters. By positing absurdly large needs, the industry is softening the regulatory terrain for faster permitting and subsidised financing of new electric capacity. Carbon hegemony reinforces this strategy – ‘drill baby drill’ (a ‘common sense’ slogan adopted by any number of recent US presidential candidates, including Trump) argues for unfettered development of fossil fuels to fill exactly such needs. AI itself has been posited as hegemony-reinforcing, adding to the exploitation of disadvantaged groups. Science Po’s sociologist, Jen Schradie suggests that elites have a distinct lead in their ability to participate, interpret, and be represented in the products of AI. Large language models have inherent biases derived from the over-representation of elites in the sources they have ingested, and the uncompensated collection of those sources is itself a source of expropriation of vast intellectual property to the benefit of tech companies and investors. The very use of AI by frontline communities can, despite these contradictions, start the evolution from ‘common sense’ – a foundation within which tensions can be exposed – to ‘good sense’, more ethical and critical approaches to our present predicaments. The present terms of the environment/AI debate are a binary based on the illusion that resource constraints are a threat to progress. By insisting that AI is profitable enough to pay its own way for clean energy, deployed responsibly, climate justice communities open an alternate channel, subversive in its insistence that deep innovation be applied to social, economic and environmental impacts as well. By deploying AI technology explicitly to build power and political will, these communities expose the ways in which the use of AI is political, create rifts in the dominant narratives, and directly challenge the digital divide. Environmental and climate justice communities today are just beginning to use AI for what civil rights leader John Lewis called ‘good trouble’. As their use and authorship of AI systems grow they will likely create hybrid approaches that unpack prior biases in AI and infuse its use with principles of democratic organising (see, for example, Jemez Principles1). Most critically, they can open new venues in which oppressive ‘common sense’ can be challenged and replaced with an organic intelligence that ‘understands, generates, acts’ counter-hegemonically. Existential threats like climate change and AI are, by their nature, windows into the contradictions that preserve an oppressive status quo. Engagement with those threats and with those contradictions are key to the emergence of ‘good sense’, shared understandings that work in favour of a more economically and environmentally just world. **References:** * Allen, T. and Coffin, M. 2022. Unburnable Carbon: Ten Years On, Carbon Tracker Initiative, * Black, S., Parry, I. and Vernon-Lin, N. 2023. _Fossil Fuel Subsidies Surged to Record $7 Trillion_, International Monetary Fund, * Chow, A. R. 2025. Inside the Memphis Community Battling Elon Musk’s xAI. TIME, * Collins, C., Flannery, H. and DeVaan, B. 2024. _Fossil Fuel Philanthropy_, Institute for Policy Studies, * Crownhart, C. 2025. Climate Change and Energy: These Four Charts Sum Up the State of AI and Energy, _MIT Technology Review_, * Gramsci, A. \[1948\] 1971. _Selections from the Prison Notebooks of Antonio Gramsci_. ElecBook, * Prabhu, S., Anthikat Suresh, K., Manal, S., Sharma, D. and Chitale, V. 2025. How Extreme Heat is Impacting India. Council on Energy, Environment and Water, * Schradie, J. 2024. Le Grand Ecart de l’IA Generative, _Comprendre Son Temps_. **_This article was originally published in Conférence issue No. 4, titled "Facing the Environmental Challenge", a publication that sheds light on major contemporary issues and informs public and private decision makers._** ### Thématique `#Environnement` `#Numérique` **Langue :** `#Anglais` --- ### Navigation pour IA - [Index de tous les contenus](https://conference.sciencespo.fr/llms.txt) - [Plan du site (Sitemap)](https://conference.sciencespo.fr/sitemap.xml) - [Retour à l'accueil](https://conference.sciencespo.fr/)