Why AI is the biggest threat to sustainability — and what you can do about it
AI is hailed as a game-changer for sustainability and climate action, but the reality is far less hopeful. Beyond its swelling carbon footprint, AI diverts attention, budgets, and energy away from sustainability while reinforcing the very systems that drive ecological breakdown. Unless challenged, reshaped, and resisted, AI risks becoming not a solution but the biggest obstacle standing in the way of meaningful progress on sustainability and climate.
In 2021, Sam Altman argued that the AI revolution was unstoppable. His outlook was decidedly optimistic: “this revolution will create phenomenal wealth,” he wrote, framing AI as a rare opportunity to redesign our world into “a much fairer, happier, and more prosperous society.” He ended his piece on an exciting note — “The future can be almost unimaginably great.”
On one point, Altman was right: AI is indeed unstoppable. Much like the rise of the Internet decades ago, I doubt its spread can be slowed, let alone reversed. But unlike Altman, I see AI less as a promise of prosperity and more as a dark cloud hanging above us, with the potential to hinder rather than support a just and flourishing future. The reason is simple: AI has become the single greatest threat to sustainability and climate action. Because sustainability and climate stability are essential foundations for human flourishing, AI, for now, is not a driver of humanity’s future, but its most dangerous barrier.
Is this about AI’s growing environmental footprint, which is already stifling tech companies’ net-zero progress? Well, that is one issue, but beyond its increased footprint, there are two more key reasons why I see AI as the biggest threat to sustainability and climate action (in short, sustainability). The first is that AI crowds out the attention, excitement, and budgets sustainability urgently needs, and the second is its role in reinforcing and sustaining unsustainable systems.
Before I explain each reason in more detail, I want to make it clear that I’m no Luddite — certainly not when it comes to AI, which I use regularly, including to edit this article and create the image above. I’m all in favor of technological advancement. At the same time, it’s essential to view our increasingly AI-driven reality not only through the hyper-excited lens of financial markets and those profiting from it, but also through a sober perspective — one that recognizes AI as a tool with multiple possible uses and impacts, not a magic wand.
And yes, I recognize that AI can certainly offer potential benefits for sustainability and climate action — from efficiencies in material extraction and industrial production to innovations tackling plastic waste and climate skepticism. Nevertheless, this positive value must be kept in perspective. It’s like reading the sustainability reports of companies such as Chevron, JBS, or JPMorgan Chase, where they highlight all the great things they claim to do to fight climate change, and yet in reality they remain part of the problem, not the solution, because their overall impact remains overwhelmingly negative.
So, here are my reasons in some more detail:
1. AI diverts attention, excitement, and budgets from sustainability and climate action
Let me ask you this — if you’re at a party and want to start a conversation with someone, what would you rather talk about: how you love/hate using vibe coding, or the backlash against sustainability in business? I assume that unless it’s during Climate Week, most people will choose the former, seeing the latter as a party-mode killer. And they’re not wrong. AI is like the new cool kid on the block that everyone is excited about, whereas I doubt sustainability — or even climate action more specifically — has ever had that status, aside from maybe 2018–2019, when the climate crisis had its heyday with activism, policymaking, and companies rolling out net-zero plans.
Does it matter that AI is cooler and perceived to be more exciting than sustainability? As Herbert Simon put it, attention is ““the bottleneck of human thought” that limits both what we can perceive in stimulating environments and what we can do.” With AI dominating the minds of C-Suites, less attention is left for sustainability. The consequence is clear: the already urgent fight against the climate crisis risks being diluted, with fewer resources and less focus directed toward what truly needs to get done.
CEOs seem especially susceptible to drinking the AI Kool-Aid when it comes to its potential for their companies. A recent Gartner survey found out that 77% of CEOs believe “AI is ushering in a new business era.” That’s not surprising given the ecstatic environment telling them that whatever they’re doing with AI is not enough. As BCG frames it, “CEOs aren’t thinking big enough with AI.” After all, it’s easy to dream big when BCG’s Jessica Apotheker calls AI “our electricity moment,” suggesting it could make business 100x more effective, the way electricity transformed industries.
Unsurprisingly, what this hype produces is FOMO. Many CEOs feel it’s riskier to wait on the sidelines than to jump into the AI pool — no matter how little they understand what’s in it. As Prof. Rene Brown, who advises numerous CEOs, puts it in an interview with the New York Times: “It looks like a complete shit show. What it looks like is scarcity. We’re not doing enough, we don’t know enough, we don’t have enough people trained, we’re not investing enough. This is what everyone’s doing and we’re behind. So it looks like fear and scarcity driving huge investments in AI that are not even aligned with business strategy.”
The surge in spending isn’t limited to big tech players like Meta, Amazon, or Microsoft, which are pouring hundreds of billions into data centers. Regular companies are also ratcheting up their AI budgets. A BCG survey found that firms with revenues above $500 million are allocating around 5% of revenue to AI, while Goldman Sachs estimates that “mega tech firms, corporations, and utilities are set to spend around $1 trillion on capital expenditures in the coming years to support AI.” Yet despite this frenzy, evidence that such investments are delivering meaningful returns remains limited. An MIT report recently made waves with its finding that 95% of investments in generative AI have produced zero returns. And as the New York Times observed, the current state reflects a “generative AI paradox”: nearly eight in ten companies report experimenting with GenAI, but just as many say they’ve seen no meaningful bottom-line impact.
So what does this surge in AI spending mean for sustainability? First, whether we like it or not, this is a zero-sum game: every dollar poured into AI infrastructure and projects is a dollar not invested in sustainability infrastructure and projects. Companies operate with scarce resources, and allocating them to AI adoption necessarily means fewer resources for sustainability adoption. Second, both AI and sustainability are inherently risky endeavors. The level of risk exposure companies are already taking on with AI suggests that we will see many failures — and that, in turn, may dampen their appetite for taking on other big risks, including those required to advance sustainability.
The focus on AI as the new engine of business transformation highlights how sustainability has lost momentum in that role (remember the “good old days” when the circular economy was touted as the path to “unlock $4.5 trillion in growth”?). At the same time, it exposes the hollowness of the so-called business case for sustainability. For years, this “business case” was cited as a barrier to ambitious action, yet the ease with which companies are now pouring billions into AI — driven largely by hype and FOMO — shows that ROI was never the real issue in the case of sustainability. Instead, the business case served as a convenient excuse to avoid changes that threatened existing (unsustainable yet profitable) business models.
2. AI is reinforcing and sustaining unsustainable systems
In a 2020 paper, The Illusory Promise of Stakeholder Governance, Lucian Bebchuk and Roberto Tallarita made the case that stakeholder governance (or stakeholderism as they call it) offers a false promise: it risks harming stakeholders by reducing managerial accountability and by delaying or diluting necessary policy reforms. They suggest that “the illusory promise of stakeholderism should not be allowed to advance a managerialist agenda and to obscure the critical need for external interventions to protect stakeholders via legislation, regulation, and policy design.”
In many ways, AI — or AIism — presents a similar illusion, carrying similar risks. Like stakeholder capitalism, it is framed as a critical catalyst for sustainable change. When Amy Luers, Microsoft’s Director for Sustainability Science and Innovation, writes that “without artificial-intelligence technologies, balancing human-caused greenhouse-gas emissions with carbon removals by 2050 is out of reach,” it echoes Klaus Schwab’s 2019 assertion that stakeholder capitalism “offers the best opportunity to tackle today’s environmental and social challenges.” In both cases, we see a familiar capitalist narrative at work: urging us to believe that the very companies fueling ecological and social crises can also deliver win-win solutions, supposedly grounded in innovation and corporate goodwill.
In that sense, AI becomes the best thing that could have happened to those promoting sustainability-as-usual, giving them yet another way to avoid the deeper redesign of business models needed to align our economic activity with planetary limits. It hands the proponents of green growth — the belief that economic growth can be decoupled from environmental impacts, long the dominant narrative in corporate sustainability — a new, shiny tool to keep that dream alive. As the World Economic Forum proclaims: “AI offers the means to accelerate progress toward halving global emissions by 2030. The task is daunting but achievable; with AI as a catalyst for scalable, meaningful change, businesses can align economic growth with environmental stewardship.”
In other words, at a moment when sufficiency should guide corporate action, AI allows efficiency to stay firmly in the driver’s seat. The difference is crucial: efficiency makes existing systems run faster and leaner, but sufficiency asks us to redesign those systems so they operate within planetary limits. Sufficiency is not about doing the same things more eco-efficiently, but doing less where less is needed, shifting practices, and embracing limits as the basis for long-term resilience. By championing efficiency, AI risks locking us deeper into unsustainable models instead of opening the door to the systemic redesigns sufficiency requires.
One can argue that this picture is overly harsh and that AI could indeed be a gamechanger for climate action. Why shouldn’t we believe that “its ability to enhance, optimize, and reinvent systems, and accelerate discovery and innovation, can help align the global economy with net-zero goals”? While I won’t deny the possibility, I find it improbable. At the end of the day, AI is just a tool, and like any tool, its use is determined by those who use it. Unfortunately, the dominant mental model shaping AI deployment prioritizes profit maximization and growth above all else. Within that mindset, AI is likely to deliver short-term efficiency tweaks rather than the deeper redesigns required to confront unsustainability head-on.
Believing in AI as a force for good ultimately means believing in the business leaders driving it. But just as we should have been skeptical of the CEOs who signed the 2019 Business Roundtable statement claiming to “redefine the purpose of a corporation,” we should take today’s talk of AI for climate with a grain of salt. If anything, the behavior of the tech executives leading the AI boom — demonstrating spineless prioritization of profits over planetary and social considerations — offers little reason to trust that AI will be steered in the direction we urgently need.
Ultimately, AI is first and foremost a business driver, and it will steer business in whatever direction those at the wheel choose. Given that business’s comfort zone lies somewhere between business-as-usual (fossil fuels, industrial meat) and sustainability-as-usual (the Amazons, Apples, and Coca-Colas of the world), it is highly unlikely that AI will shift that range. If anything, it will be used to reinforce and extend this comfort zone. Like every technological advance before it, placing too much faith in AI’s promise means willfully ignoring how systems work — and the kind of change the climate crisis truly demands.
So, what can we do?
The threat AI poses to sustainability should be met with action, not passive acceptance of its current unsustainable path as a force majeure. I see three main ways to push back:
Challenge the narrative
we need to confront the dominant story that frames AI as a force for good in climate action. Too often, the debate is reduced to the carbon footprint of data centers or the energy cost of personal AI use. Important as those are, focusing only on this part misses the bigger picture — and the deeper threats AI poses to sustainability. As Pete Bronski argued, we must “stop buying into the climate promises of AI-powered solutions hook, line, and sinker. AI currently sits at a precarious point on Gartner’s Hype Cycle, and it’s important to separate true solutions from hollow promises.” I couldn’t agree more. We should not be afraid to go against the fuzzy consensus around AI, and we must raise questions and expose some of the less visible potential threats it poses.
Change the environment (dark matter)
If we assume that AI, like anything else in business, is driven by the environment in which it operates, then it is more effective to focus on changing that environment — regulation, social norms, financial mechanisms, and culture — rather than only on AI implementation itself. This means working on what Dan Hill describes as the “dark matter”: the often invisible forces and infrastructures that shape how systems (companies included) function. Engaging with this dark matter can take many forms — policymaking, activism, norm entrepreneurship, communication, innovation, design, and more. The point is to create favorable conditions for AI to evolve in ways that support sustainability, and unfavorable conditions for it to develop in the opposite direction.
Resist
Holly and Will Alpine have left their work at Microsoft to protest the company’s work with oil and gas companies aimed at “automating and accelerating oil extraction.” In response, they launched the Enabled Emissions campaign, which focuses on “holding Big Tech accountable for accelerating fossil fuel production.” As they point out, AI’s role extends far beyond data centers: “While AI’s direct carbon footprint is gaining significant attention, its deliberate role helping oil companies significantly increase fossil fuel expansion — with staggering emissions — remains largely unaddressed.” Their campaign is a vivid example of how challenging the narrative can evolve into active resistance. Such resistance — whether inside or outside organizations — is essential. Employees, activists, and stakeholders all have a role to play in demanding not just accountability but genuine alignment with net-zero goals. If AI is being weaponized to prop up fossil fuels and other unsustainable activities, then directly challenging companies on these practices can make them less willing to pursue them.
Raz Godelnik is an Associate Professor of Strategic Design and Management at Parsons School of Design — The New School. He is the author of Rethinking Corporate Sustainability in the Era of Climate Crisis. You can follow me on LinkedIn.
