A few centuries ago, much of the world’s natural systems lay outside the reach of human activity. Today, direct and indirect impacts of our actions touch every ecosystem, from glaciers and oceans to savannahs and forests. Every ton of CO₂ emitted, and every year we delay acting, shapes the planet as surely as a deliberate intervention would. In a world where human influence is already impacting every ecosystem, to refuse to act is still to act; to avoid intervention is not a neutral action, it simply lets existing interventions run their course1. The question is not whether to intervene in Earth’s systems, but how to do so responsibly.
A world that hits net zero in 2050 will see rising temperatures for at least the next 25 years. That means at least a generation’s worth of time, where the climate today is the best there is. It means more extreme weather events, more storms, more floods, more droughts, more human suffering and more economic damage – and that’s in a world where we manage to decarbonise on timescales that are looking increasingly unlikely.
There are other cases where we hit tipping points along the way that make it even more difficult to stabilise the climate. The level of certainty around these scenarios is variable, but one thing is true: the warmer the planet gets, the more likely we are to reach these points.
What are our options? There are ideas, such as stratospheric aerosol injection, glacial stabilization, marine cloud brightening, and others, but most geoengineering approaches remain at the conceptual or early research stage. Their immaturity is often cited as reason enough to avoid them, as though ignorance were a form of caution. Some have concerns that further work on these options will at best distract and at worst act against the push for decarbonisation. While it’s important to consider this risk, ignorance about the potential of these options is not a neutral position. We may one day face conditions that make some form of intervention unavoidable.
However, researching these approaches does not commit us to using them. The purpose of early stage research and trials should be outcome neutral; the same experiments and modelling that might one day enable deployment could also reveal why these interventions shouldn’t be used. More knowledge can be a safeguard as much as an enabler, and refusing to know leaves us blind in both directions.
Hazards
There are good reasons to hesitate before pursuing the kinds of interventions listed above. Some argue that even research into them will create various moral hazards, slowing the political and economic drive toward decarbonisation or adaptation. These concerns deserve attention, but we should be clear about what kind of claims we are making about these interventions because different views suggest different risks of unintended consequences.
In many critiques, the concern is that geoengineering will become the focus of climate attention and will be viewed as a solution to a warming world. Yet in almost every conversation I have had with people in the space they simply want us to have better understandings of these approaches, not as solutions but as contingency plans. The aim of the research these people advocate for is not to normalise the use of these techniques but to map their contours, so that if we ever reach a moment of necessity we have a more complete picture of the options available to us.
The primary concern people have about these interventions is mitigation deterrence, the fear that geoengineering could slow emissions reductions. That might happen indirectly, if funding and attention shift away from decarbonisation, or more deliberately, if powerful groups use these approaches to create “predatory delay.” This concern is not hypothetical. Technologies like CCS or CDR were used in integrated assessment models long before either reached commercial maturity, with the result that the level of pure decarbonisation ambition was lowered in these modeling results, and therefore in most IPCC projections2.
In principle, if there are politically feasible and technically proven ways to cut emissions, then time and money spent on geoengineering could be a distraction. But we should also be clear-eyed about the world we live in. After three decades of COPs, the Keeling Curve hasn’t bent. Betting everything on a set of technologies and political willpower that has not managed to deliver in practice is not obviously the cautious choice3.
The risk of predatory delay also remains real. Some fossil-fuel interests have historically advocated for green technologies or policies as cover to slow genuine transition e.g certain oil companies have championed carbon pricing precisely because they thought it would be low ambition and therefore posed minimal threat to their business. Any work in this space wants to try to ensure that it isn’t used as a substitute for ambition, especially at a stage when the efficacy of these approaches is still being determined.
To return to the theme of the opening paragraphs, refusing to research geoengineering approaches is not a morally neutral act if that ignorance leads to more heatwaves, crop failures, and ecosystem loss. Still, we should be very careful with this line of argument. It is easy to slide from “avoiding research could worsen suffering” or “we may need these approaches as an emergency option” into a blank cheque for irresponsible research and trials. The ethical terrain here is steep: the same logic that justifies preparedness can also be misused to license premature deployment. This is why the process of research and trials will, for the foreseeable future, be just as critical as the product of these efforts.
A second, less discussed hazard is adaptation deterrence: the risk that research and funding for climate interventions substitute for work that helps communities directly adapt to the changing climate. In many ways geoengineering is more of a substitute for adaptation than mitigation. Cooling the planet through solar radiation management or working on glacial stabilisation addresses the same type of problems as drought-resistant crops or flood defences just at a much more general scale. If money and attention are finite; a dollar spent on geoengineering research may be a dollar not spent on local resilience.
Whether geoengineering as a method of adaptation is harmful depends on how we frame these interventions. We already accept coping strategies when we build sea walls or breed more resilient grains. Those interventions are celebrated because they directly reduce suffering often for the world’s most vulnerable.
Geoengineering has often been advanced under the same moral banner, protecting the world’s most vulnerable. We need to be careful that those communities have a say in the work done in their name. It’s easy to imagine that “more geoengineering R&D” would not top the wish list of those most exposed to climate change. If the motive of helping these groups is to carry ethical weight, it must come with participation; it would be wrong to use people on the frontlines of climate change as the justification for this research while not having them involved in the agenda setting and knowledge production itself
There is another adaptation lens I want to touch on. Traditional adaptation measures and climate interventions differ not just in scale but in kind. Most geoengineering proposals act as generalised adaptive interventions: if you brighten clouds above the Great Barrier Reef, the entire ecosystem benefits more or less equally. By contrast, conventional adaptation is necessarily selective. Building sea walls, irrigation systems, or drought-resistant crops requires deciding which places, species, and communities receive protection and which do not.
Seen this way, climate interventions appeal to a more naturalistic ethic, one that aims to preserve or stabilise the conditions that allow ecosystems themselves to endure, rather than focusing primarily on human welfare. Humanity has always adapted to change. The same cannot be said for rainforests, coral reefs, or glaciers. We can design crops that survive drought, but there is no straightforward analogue for saving the Amazon or the ancient glaciers that have accumulated over millennia. Yes, there are things like conservation areas but without addressing the driving climatic forces, adapting these systems to a new world will remain a huge challenge.
Governance and safeguards
I’m deliberately not going to spend many lines on the question of governance. Not because it is unimportant, the opposite is true, but because I want to address more of my thoughts on the topic in a future piece. What I will say is that while all the climate inventions I’ve mentioned remain in the realm of speculative technologies, we should focus our research and governance structures less around propelling any given approach to TRL10 and more around the process of knowledge generation itself. Research in this space is inevitably contentious, but without a deliberate process, the knowledge we generate seems more likely to inflame existing disagreements than to reduce them4.
We want broad participation in these efforts, and ideally, rules and structures to prevent overuse and overreliance on these approaches. Alongside this, we should be building our understanding of how these approaches work and the institutions to help us deploy them, if needed, in a way that is not a knee-jerk reaction but a considered approach.
We no longer stand apart from the systems around us. Every action we take, or refuse to take, leaves an impact. Geoengineering forces us to face that fact directly: human influence over the planet is not something we can renounce, only choose how to exercise. The task is not to preserve a fiction of non-intervention, but to learn how to act within a world already shaped by us. Climate interventions are, at their core, delay tactics, but in a world of destabilising climate, we may well find ourselves wanting to buy some time.
- My friend Peter has an example in a post he wrote that made me stop in place for a bit that is along the same lines so I’m just going to quote the whole thing “For example, what happens if some geoengineering experiment goes wrong, and say, like half of the insects on earth die? Never mind! We’re already doing that”. ↩︎
- This is largely a feature of how these models work, IAMs look for the lowest cost way of reaching certain temperature targets. If it turns out that CDR is cheaper than fully decarbonising aviation then CDR will show up in these models. Viewed in this way this form of potential “mitigation deterrence” is instead a much more reasonable adjustment to the enormous cost implications of trying to directly decarbonise hard to abate sectors. Whether this directly changed what countries did is debatable, most have missed their own targets regardless, but it shaped the sense of what counted as “feasible”. ↩︎
- I am very sympathetic to the view that we are just now reaching the rapid growth phase for technologies like solar so we should in fact expect the future to look drastically different to the past even without much more policy support. However, whether this will be enough remains to be seen. ↩︎
- This is something I feel uncertain about. Designing the process of knowledge generation with care seems essential if any resulting knowledge wants to carry legitimacy rather than increase polarisation. But given the political terrain of geoengineering, I’m not sure how far process design can really move us toward consensus. Maybe the more effective path to legitimacy is not in how knowledge is produced, but in how openly and inclusively it’s communicated once produced. ↩︎
Neil Hacker 