Hide table of contents


  • This post analyzes whether phytomining, a process where plants are used to extract valuable minerals, would be a good technology for people affiliated with Effective Altruism to promote.
  • It analyzes how phytomining could further Effective Altruism goals and explores how different cause area prioritization decisions and moral issues may contribute to that
  • It makes some preliminary suggestions about further steps that Effective Altruism organizations could take to explore if supporting phytomining would meet cost-effectiveness frameworks.
  • I am not an expert on anything even vaguely related to phytomining. This means that everything should be taken as very tentative and as a suggestion for further exploration rather than a final conclusion.
  • It concludes that phytomining is especially promising for further investigation in civilizational resilience, but also is worthy of significant consideration for its impacts on global development and also climate change.
  • It suggests that phytomining has positive but less significant effects on global health and biodiversity.
  • It notes the possibility of unintentional negative effects on nuclear risk, artificial intelligence, and animal welfare.
  • It notes that it is very promising for natural resource depletion, but that that is not an area of high concern for Effective Altruism.


This post is long. If you don’t want to read it all the way through, I recommend reading this summary, skimming the first part of the theoretical explanation and maybe the practical explanation, reading the cause areas that interest you the most, and then reading the analysis and maybe the next steps.

Positionality Statement:

I’m not formally affiliated with Effective Altruism nor do I consider myself part of the Effective Altruism movement, although I do consider myself something of a fellow-traveler with respect to many of the movement’s goals. Again, I have no expertise in anything related to phytomining (geology and soil science, botany, the economics of mining). I’ve put a sizable amount of effort into verifying that all facts in this post are accurate, but unfortunately it’s possible some inaccuracies may have slipped through. Please let me know if you find any inaccuracies.

All conclusions that I draw in this post should be considered to be very tentative. I’ve spent very roughly about 15 hours on this post, but I could easily foresee spending additional time on the topic (especially if it involved talking with experts) leading to me changing my mind with respect to the benefits of phytomining. This post is intended to encourage further exploration of the relationship between phytomining and Effective Altruism. It’s not intended to be the final word on phytomining, and I would strongly advise against anybody else taking actions because of this post without doing some research themselves first. I began to work on it as a standalone post, changed it up to submit it to the CEARCH cause area contest, and then made slight modifications to that version for the purpose of posting on the Forum.

I’ve used in-line citations for all sources that aren’t van der Ent et al. (2021), because that book was my primary source for this entire post and it would have grown too unwieldy to cite it every time I referenced it. Additionally, some people make a distinction between phytomining (usually as the specific production of metal from a plant) and agromining (the entire system of producing the plants, getting the metal from them, and everything else involved with that), but I’ve chosen to refer to both of them as phytomining given that that seems to be a common convention in news reports, and because I personally find the distinction a bit confusing.

This post was entirely written and researched by me without the use of ChatGPT; I later tried to use ChatGPT to proofread it but it gave very few usable suggestions. Thanks to some friends who agreed to read this over; all errors contained herein are solely my fault and not theirs.

Explanation of Phytomining:


The core of phytomining is that it is a process that takes plants (and soil and soil amendments) and turns them into metal (and some waste products). However, the process of phytomining starts long before the metal is extracted from the plants. Really, it started with the beginning of the universe, but let’s cut to the chase and look at when human involvement begins to take place within the phytomining process. Botanists typically lead the way in identifying species of plants that are hyperaccumulators, which means that they are capable of taking in much more of their target metal from the soil than a normal plant is able to. The exact thresholds vary significantly depending on the author and the metal, but usually the concentration of the metal in plant tissues of a hyperaccumulator is a few orders of magnitude higher than it is in the plant tissues of a sampling of average plants. 

However, there are many complications in identifying hyperaccumulators, besides the ever-present risk of measurement error. To start, the concentration of a metal will differ drastically between burnt plant material and unburnt plant material, and both are sometimes reported. Although most research into hyperaccumulation has focused on leaves and stems, other parts of plants are sometimes known to have particularly high concentrations of target metals, and the concentration of a specific metal may vary depending on which part of the plant is measured. An additional complication is that the extent to which a plant will hyperaccumulate does depend on the inherent concentration of the metal in the soil. Some plants only engage in hyperaccumulation if the metal concentration in the soil is already high, while others will have significantly higher accumulation than other plants even on soil with low concentrations of the target metal. Furthermore, it also varies whether hyperaccumulator plants grow only on soil with a particularly high concentration of the target metal or are also found on other soils. Good hyperaccumulators are usually perennials that grow quickly and widely, are easy to propagate, and collect high concentrations of the target metal.

This is a good place to take a look at what kinds of soils contain the target metals necessary for a successful phytomining operation. Typically, ultramafic soils are known for having high concentrations of metals. I’ll talk more about this in the practical explanation, but ultramafic soils make up about 1-3% of Earth’s land surface, depending on which estimate you go with, and are distributed worldwide, albeit with particularly notable concentrations in the Mediterranean and the tropics (Kidd et al. 2018). Yet phytomining is not limited to these naturally occurring metal-rich soils. Soils that have been contaminated by metals are also a promising avenue for phytomining operations, as are the tailings of mining operations. Research is ongoing as to whether phytomining can be used as a form of metal recycling, with metallic waste as a component of the soil. Hydroponics are sometimes used to test metal intake in experiments, but it doesn’t seem to me like hydroponic phytomining has known uses in the real world.

Still, simply knowing that hyperaccumulating plants are likely to be found on ultramafic soils is not enough to easily narrow down the question of which plants are hyperaccumulators and which ones are not. It often takes a significant amount of labor by botanists to identify plants that engage in hyperaccumulation. Traditional tests to identify hyperaccumulators are often difficult to perform in the field. However, the rise of portable X-ray fluorescence technology has recently permitted easier identification of hyperaccumulators. These devices, which can analyze the metallic composition of plants, have already been used successfully to identify many likely hyperaccumulators from collections of plant samples. 

Even after a given hyperaccumulator has been identified, there’s still a lot of testing necessary to identify how to use a specific hyperaccumulator most efficiently. A major component of this testing is work on how the plant grows best. Many hyperaccumulator plants benefit from some form of fertilization or other soil amendments. Some plants are also tested to examine how quickly they grow and how often the parts of the plant that contain the metal (usually this is leaves and sometimes branches, but for annuals it can be the entire plant) can be harvested. 

Testing can also reveal how best to extract the metal within the plants. A lot of the details about this, and, frankly, testing as a whole are fairly technical and don’t seem particularly relevant for the audience of this post, so I didn’t spend as much time examining this stage as I did some of the other ones. However, it’s clear that there are a wide range of options for extracting the metal, depending on both the kind of plant and the kind of metal, and that this can have significant impacts on the climate analysis later on. A common technique is to burn the plant matter first to concentrate the metal, which might be combined with leaching in order to separate the metal from the waste material. Other approaches involve trying to leach the metal directly from the plant, but this strategy is newer and is still being developed. 

After the phytomining potential of a plant and soil region is established and some kind of tentative advice is known, there are hopefully larger-scale tests as to whether the phytomining procedure works well in the real world. These harvests can scale up into full-fledged commercialization, where farmers either sell their crop to some kind of smelting or metal extraction facility or possibly operate such a facility themselves. I’m working off of very little data here because of how little phytomining has made it to this stage of operation, but it seems likely that scientists will become less important in the operation of the facility and more of the decisions about mining will transition to locals, specifically the people who are operating the facility. However, this probably does not mean that the phytomining process is no longer undergoing improvements, but rather that decisions are being driven by commercial rather than scientific considerations. This is the end goal of any phytomining development process.


However, there’s still a large question remaining: what the current state of the field of phytomining is. With regards to the first step, many hyperaccumulating plants have already been identified since the idea of phytomining first gained some currency towards the end of the 20th century. As of van der Ent et al. (2021), the Global Hyperaccumulator Database of the Center for Mined Land Rehabilitation at the University of Queensland listed 721 different species known to hyperaccumulate a wide variety of metals; doubtlessly X-ray fluorescence has allowed many more to be identified since then.

These plants additionally represent a wide variety of ecosystems. Nickel hyperaccumulators have been extensively studied in the French overseas territory of New Caledonia, located in the southwest Pacific Ocean not too far from Australia. The Malay archipelago (Malaysia, Indonesia, and the Philippines) has also been the site of analyses of nickel hyperaccumulation. Furthermore, the Mediterranean region of Europe, such as the Balkans and Italy, is also known to contain a good number of species that can hyperaccumulate nickel. Other nickel hyperaccumulators have been identified in Cuba, South Africa, and a few different parts of Mexico, including one not on ultramafic soils. Copper hyperaccumulation has been identified within the Democratic Republic of the Congo. Selenium hyperaccumulators have been identified in the continental United States of America and China. And of course this list is far from comprehensive. This probably shouldn’t be surprising, but I’ve noticed something of a trend in these examples of hyperaccumulators being identified in regions already known to produce the target metal.

As you might have guessed from the above section, the metal with the most research on its hyperaccumulation is far and away nickel. Commercial nickel endeavours are already taking place in Europe as part of the LIFE-AGROMINE initiative and through the French company Econick (Alchemia Nova n.d.). I’m not entirely confident on how either of them are doing, but they both do appear to still be in operation, and Econick apparently raised some additional later-stage venture capital in late 2022 (Pitchbook n.d.). Another European phytomining concern called Stratoz doesn’t seem to have any current information about them on the Internet. There are also large-scale tests that are moving towards commercialization in Sabah, the Malaysian side of Borneo (Morse 2020).

Other metals have not seen quite the same enthusiasm as nickel, but that does not mean that research on their phytomining possibilities are entirely lacking. Thallium, a toxic metal that is important for many electronics, has been flagged as potentially quite promising for phytomining because it accumulates very easily in plants and thus could be easy to turn a profit. Cobalt has also been judged promising. Gold and other noble metals have been examined for their phytomining potential, but the reviews in van der Ent et al. (2021) are not particularly positive. Other research is ongoing concerning copper, manganese, selenium, and rare earth metals. There are more speculative reports of hyperaccumulation and thus the possibility of phytomining for a wide variety of metals, including mercury, technetium, uranium, and thorium. 

For many other metals, research has been driven by the related topic of phytoremediation, where plants are used to clean up hazardous, typically contaminated, areas. Phytoremediation can, however, be combined with phytomining to both produce products of direct economic value and soils that are suitable for further production. Cadmium enjoys a high demand for its phytoremediation, but its prospects for phytomining seem worse because of a lack of efficient hyperaccumulators. Certain fern species have been used to phytoremediate arsenic, but it’s not clear if the demand for arsenic is high enough to make the phytomining of arsenic a particularly likely outcome. Selenium is another metal where phytoremediation and bioextraction work may eventually lead to its phytomining in the United States and China.

It’s very hard to find a solid estimate for how many people are working on phytomining currently. Probably the largest single organization working on phytomining as a major component of its mission is the Center for Mined Land Rehabilitation at the Sustainable Minerals Institute of the University of Queensland (University of Queensland n.d.). This is partially driven by the fact that a lot of phytomining work includes crossover with phytoremediation and other elements of research on the relationships between metals and plants. Furthermore, many phytomining researchers are academics and so also devote some of their time to teaching and other topics separate from their research. There are also chapters of van der Ent et. al (2021) that reference master’s theses, but it’s unclear that those individuals, or, for that matter, anybody whose current research I did not explicitly look up stayed within the field of phytomining research after that. There are ~60 contributors to van der Ent et. al (2021). Names in other papers that I looked at match very closely with the list of contributors to that work. If I assume that everybody who works on phytomining research contributed to the book and that they each spend half their work on phytomining, that results in an estimate of 30 FTEs (full-time equivalencies), the vast majority of which is devoted to scientific research and a smaller slice devoted to commercialization. 

There clearly is some work ongoing towards phytomining policy and the non-scientific development of phytomining as a plausible alternative and complement to conventional mining. For instance, phytomining was mentioned as part of a law passed by the French government concerning the circular economy. Furthermore, the Department of Energy, through ARPA, has funded research concerning phytomining in the United States (Krol-Sinclair and Hale, 2023). However, it’s not really clear to me if anybody is devoting a substantial proportion of their work time to focus on policy incentives to scale up phytomining production. There are also, of course, certainly some workers who tend to the plants on phytomining farms and don’t directly do research, but I have no idea how many there are and it doesn’t seem decision-relevant.

Global Health and Development Impacts:

One of the major cause areas in Effective Altruism is global health and development, although it can also be divided into two separate cause areas of global health and global development. Phytomining can be judged as an effective technology to improve both global health and global development. For global health, phytomining is useful because it serves as a substitute for conventional mining, which tends to produce poor health outcomes in both miners and the surrounding community. For global development, phytomining is useful because the jobs associated with a phytomining operation can often be preferable to jobs in traditional mining. However, support for phytomining because of its impacts on both global health and global development is indicative of a view among Effective Altruism that the near-term future should be the top priority, so these causes have been grouped together.

Global Health:

There aren’t great statistics on how much of a health risk mining poses. According to the CDC, there were 37 known occupational mining fatalities in the United States in 2021, the last year for which data has been reported (Centers for Disease Control and Prevention n.d.). This is equivalent to 16.1 fatalities per 100,000 full-time employees. However, the United States is clearly capable of having much stronger regulations on mining than the developing countries where a large amount of metal production occurs. Furthermore, the health issues associated with mining are not just ones linked to current miners but also diseases such as black lung and the impacts of pollution and contamination from the mines on the surrounding community. The scale of issues with pollution from mining facilities, particularly upper respiratory infections, has been reported to be quite high in certain areas (Timmerman 2022). Yet it’s also important to take into account that the health impacts of mining differ not just by the region but also based on what is mined; for example, uranium mining involves exposure to radon (Stephens and Ahern 2001). Phytomining can replace conventional mining for many common metals such as nickel, but it cannot replace coal mining.

One challenge in establishing the impacts of phytomining on global health is that phytomining does not remove all of the negative health outcomes linked to mining. Many phytomining procedures still involve some degree of smelting, although phytomining often results in higher-quality ore so less smelting can be necessary, and there are attempts to develop procedures that eliminate or cut out entirely the use of smelting. Thus the reduction in air pollution from implementing phytomining instead of conventional mining is variable but seems to generally be positive, although I’ll flag this as an important point for further research and consideration in terms of global health impacts. However, it’s safe to say that phytomining has much fewer of the health impacts of working underground, since if you’re trying to grow plants underground you’re almost certainly doing it wrong. 

In the end, I wouldn’t rate phytomining as a particularly promising avenue for further exploration by the Effective Altruism movement based on global health considerations alone. However, I do think that the global health impacts should be taken into account when looking at phytomining principally as a cause in other areas. Mining makes life pretty terrible for the communities surrounding it, but the scale of health impacts from mining, although hard to determine, seems to be less than that of major tractable issues such as malaria (which kills 1.1% of people worldwide) and other infectious diseases (Dattani et al., 2023). 

Global Development:

Mining is a major industry globally, but again it’s very hard to say exactly how many people are employed by the mining industry worldwide, particularly because not everybody who works for a mining company is a miner. A 2019 report by the World Bank estimated that more than 40 million people work in small-scale or artisanal mining worldwide (Hobson 2019). A trade union group estimates 3.7 million people work in the mining sector worldwide, with 2.2 million of them in developing countries (IndustriALL n.d.). 

On one hand, mines have historically been able to draw workers by offering relatively steady and high-paying jobs, especially compared to the forms of seasonal employment that are common across the developing world (Timmerman 2022). On the other hand, mines often impose high negative externalities on the local community and have been accused of both contravening the will of local communities that do not want to see a mine and using inaccurate predictions of their impacts on the local community. Some examples of these negative externalities include environmental disruption that removes the option of traditional forms of employment such as farming and fishing. Both air pollution and noise pollution can also drive residents to seek employment further from their communities. 

Phytomining, by contrast, offers what I would consider to be a much rosier employment picture, although I’d caution that you particularly shouldn’t trust me, a lifelong inhabitant of a high-income country, on the preferences of people in lower-income countries. Because much of the core labor of the phytomining process is agricultural, it can fit well into the growing season of farmers, a particularly common occupation in the developing world. The metallic soils used for phytomining are typically considered not particularly good for growing other crops, so there’s a very low opportunity cost to using the land for phytomining instead of other commercial practices. The cost in both money and effort of establishing a mine in a given location is much higher than the cost of establishing a phytomine (both need some form of a metal processing facility, but the mine also needs a lot of actual mining and the phytomine just needs some fields). This means that phytomines can be started more easily, and, because they have less fixed costs to cover, may also be stopped more easily. Phytomines also are less likely to require the displacement of people living on top of the mine, because, like other agricultural operations, they can work with non-contiguous fields.

But phytomining isn’t just good for the people who work and live near the mine. This is getting a little speculative here, but the Indonesian government has used its control of mineral wealth to strongarm companies into building processing infrastructure for metals such as nickel in Indonesia, and using that to help get more advanced manufacturing located on Indonesian soil. This strategy, which is called downstreaming, is definitely controversial and comes with significant risks because it contravenes much of the dominant free-trade logic of political economists. Phytomining can help with this because it can both make more metal and tweak the distribution of where metal is made. Even without those measures, however, there’s a common-sense argument that location-based benefits mean that it’s often a good idea to produce metal near where the metal is used. Thus having a strong mining industry can help set a country on a path towards global development. However, it’s still important to note that this section is very speculative and I don’t see it as a particularly major benefit of phytomining.

That doesn’t mean that phytomining is without its downsides from a development perspective. Mines are often located in out-of-the-way or remote areas, such as in the highest reaches of the Andes mountains. Phytomines, by contrast, need to be located in areas well-suited for the growing of crops, and are more likely to coexist with agricultural areas. Thus there are likely distributional consequences from establishing phytomines instead of traditional mines, although phytomining can often be done on former mines or areas next to present-day mines. My sense is that this mostly ends up favoring phytomines because humans have a very strong revealed preference to live near friends and family, but views could quite reasonably differ on this. Furthermore, traditional mining is much harder to ruin with bad weather than phytomining, meaning that the risk of phytomining is positively correlated with other risks from working in agriculture. 

These concerns are definitely very real, but I don’t think that they outweigh the many positives of phytomining from a development perspective. If you boil down the story of phytomining with regards to global development to its most important elements, it goes something like this: Mining employs a lot of people and can make a lot of money. But communities on the ground don’t like to adopt it because of very real downsides. Phytomining lets them get the positives of mining (a new income stream, something to attract investment in other industries) without the downsides (pollution, mostly).

Climate Change and Biodiversity Impacts:

The conventional argument in favor of phytomining, presented in many of the news articles that I read for this, heavily focuses on phytomining’s environmental impacts. For Effective Altruism’s purposes, it’s especially useful to divide these into two: effects on climate change and effects on biodiversity. Biodiversity isn’t usually considered a major cause area, but there has been some discussion of biodiversity as a possible cause for the Effective Altruism movement (Malhotra 2022). The climate change case comes from how phytomining can both increase the supply of metals useful for green technology and how it can get metal for less greenhouse gasses than conventional mining. The biodiversity case is about how it can be better integrated with natural habitats than conventional mining, which often destroys the habitats, and also how phytomining can have benefits for overall ecosystem health. My understanding is that natural resource depletion is considered very speculative and unlikely by Effective Altruism, but I’ll also mention its interactions with phytomining briefly.

Climate Change:

There’s high demand for many metals because of green technology. Nickel, one of the elements with the highest promise for phytomining, is also particularly crucial for a wide array of clean-energy technologies, including electric car batteries (Roberts 2022; Timmerman 2022). Krol-Sinclair and Hale (2023) are particularly bullish on the prospects for phytomining of cobalt — I’m not sure how much to read into this, since it’s newer than van der Ent et al. (2021), but also written by people who are not scientific experts — and cobalt is also critical for many green technologies. Lithium, another essential component of certain battery designs, likely has hyperaccumulators but is much further behind in development than nickel and cobalt. Zinc and aluminum both have somewhat better outlooks than lithium. However, there are also other important metals that don’t seem to accumulate or are too toxic to plants to be considered a good candidate for phytomining, of which chromium is an example.

The combined impact of an increased and cheapened supply of these metals could help catalyze the transition to a climate-friendly future, but the impacts of phytomining also go beyond that. Common sense (and van der Ent et. al (2021)) suggests that phytomining is more energy-efficient than traditional mining; there’s no need to dig a hole in the ground. It’s also likely that the metal can be processed more efficiently because of the form that it is in. The climate impacts of transporting the metal from a phytomining site versus a conventional mine is hard to say, but I suspect that it’s a minor consideration as opposed to energy-efficiency and supply of critical materials. However, I will flag that people’s beliefs about which technologies are most promising for future climate impact is an important consideration when judging the climate promise of phytomining. For me, I’d judge phytomining to be promising by traditional standards, but I struggle to envision a scenario in which phytomining outcompetes climate interventions currently endorsed by Effective Altruism.


The case for phytomining for biodiversity reasons is based on three connected propositions: phytomining fits well with natural habitats, it can restore soils, and it can incentivize the protection of natural habitats. The third one is probably the least significant; phytomining incentivizes the protection of natural habitats because they can be used to find new hyperaccumulator species. I don’t consider this line of reasoning particularly promising because preserving a wide range of plant species is already useful for pharmacological purposes. 

I already discussed phytoremediation, the use of plants to clean up soils, in some earlier sections, but it’s a critical element of the second point for biodiversity. One of the more promising types of locations for phytomining is contaminated soils. If these are phytomined, the level of contamination can thus be decreased, creating the possibility of the land again providing a habitat for the species that were on it. On the other hand, particularly large-scale phytomining could reduce the incidence of metal-rich soils and thus hurt the prospects for certain hyperaccumulators that rely on it. However, this doesn’t seem to me to be a particularly concerning downside because it would rely on a tremendous amount of land being devoted to phytomining and additionally ignores that many hyperaccumulators can also grow on normal soils.

This leaves the impacts of phytomining on being better than conventional mining in preserving natural habitats. On one hand, this makes sense because phytomining mostly uses soils that aren’t considered good by the normal standards of plants, while conventional mining tears up whole ecosystems in order to find the metals underneath. Still, it’s possible that further research could determine negative impacts to phytomining, since it does require changing an ecosystem with the introduction of additional hyperaccumulators for harvesting. Focusing on native hyperaccumulators, which is suggested anyways because they are likely to be well-adapted to the environment, can help remediate this concern. 

It’s hard for me to form a particularly coherent opinion as to the biodiversity impacts of phytomining, because so little research done in this area has explicitly compared it to the standards of the Effective Altruism movement. Still, I’d very loosely guess that phytomining has positive biodiversity impacts, but that there are better marginal uses of time and money in terms of promoting global biodiversity.

Natural Resource Depletion:

The level of concern allocated to natural resource depletion seems to me to generally be quite low among Effective Altruism, but it is given more serious consideration by individuals outside the Effective Altruism movement (Wiblin, 2016). Still, I think it’s worth addressing what phytomining could do to natural resource depletion. Phytomining can help restore soils because many plants do not thrive in soils that have high concentrations of the metals that are removed through phytomining, although there is also a risk of erosion. Phytomining can also increase the supply of metals that are extractable for a reasonable price. On the other hand, phytomining supplies on naturally ultramafic soils do not last forever (although presumably soil contamination and the more hypothetical industrial waste approach will continue to have opportunities), which could lull civilization into a false sense of complacency about the ease of metal supply. I’m especially ill-informed about natural resource depletion, but I’d say that phytomining could be highly effective in preventing natural resource depletion by increasing the supply of metals, but also that I’m not particularly concerned about this topic.

Civilizational Impacts:

Civilizational resilience is yet another lens through which phytomining can be judged for promise as a cause according to the principles of Effective Altruism. The theory here is that, if we’re in the unfortunate position of redoing civilization after some kind of major catastrophe or near-extinction event, it’s important to have easily accessible metals. Furthermore, many of those easily accessible metal sources have already been used in the past, so this future version of us would have a harder task than the one that our civilization had before. Phytomining can ameliorate this because it doesn’t seem to require anything that would be particularly hard to do after a collapse. There’s no need to dig huge holes, it can work with a variety of options for refining the metal, and even the identification of hyperaccumulators has been done without fancy technology. However, one possible challenge is identifying hyperaccumulators post-catastrophe; knowing what they are before a catastrophe occurs would almost certainly be easier.

One major question here is what metals would be needed in such a scenario. I don’t have access to the full book, but Dartnell (2014) speculates that the top priority would be making steel, which takes iron, and that aluminum would also be particularly useful. Neither of these metals have been the subject of much phytomining research or have been judged to be particularly promising. However, that doesn’t mean that there isn’t a need for, say, nickel, thallium, and cobalt, to name some metals where phytomining seems more promising. One important avenue for further research would be figuring out what metals are both critical to civilization and comparatively easy to phytomine. All in all, I think that phytomining could potentially be quite impactful from a civilization resilience lens and is particularly deserving of further research. 

Other Impacts:

Here I’m going to lump together three cause areas that all are very important to the Effective Altruism movement that could potentially have minor negative interactions with phytomining: nuclear safety, artificial intelligence and technological challenges at large, and animal welfare, especially with regards to wild animals. These don’t share that much in common, but it’s convenient to address them at the same time. I’ve tried to brainstorm any possible interactions with remaining cause areas (pandemics, aligning present-day technology) but can’t think of anything. 

Nuclear risk:

Let’s start with nuclear risk, where I consider phytomining’s prospective impact to likely be largest. Metals, especially uranium, are important to the production of nuclear bombs. Uranium can theoretically be produced using phytomining, although research into this is very scant compared to more popular metals like nickel. It’s possible that some other naturally occurring radioactive elements, such as technetium, could also be phytomined. However, there are reasons to think that this isn’t a major concern. One is that there are a lot of countries with nuclear power, and thus uranium, but no nuclear weapons, including countries such as Iran that would like to have nuclear weapons. This is evidence that the pure availability of raw uranium is probably not the limiting factor in the production of nuclear weapons. I looked for a more reliable source on this, but the World Nuclear Association, a pro-nuclear power trade group, also heavily implies that it’s uranium enrichment, not the availability of uranium, that limits countries (World Nuclear Association 2023). It’s also worth considering that a lot of phytomining research is specific to certain metals, and so is unlikely to significantly change the calculations with respect to uranium availability. I consider this something to keep an eye out on in future phytomining research, but it seems like a very minor and speculative downside to this technology.

Artificial intelligence:

The effects of phytomining on the development of artificial intelligence are more or less the effects of phytomining on technology development at large. Lots of metals are useful in building technology and in particular semiconductors, including rare earth metals, which have a potential for phytomining (IEEE n.d.). Electronics are also a major use case of thallium. For the most part, I think it’s widely held that cheaper technology is good. But some of this technology can be used in AI capabilities research, and training runs for large language models are expensive, although I can’t find any high-quality sources estimating what part of that is the cost of electricity versus the cost of the computers used to run it (Sajid 2023). It doesn’t seem to me, however, that the cost of the specific metals with the greatest phytomining potential is a limiting factor in these runs, however. Any impact of cheaper electronics from phytomining would likely be slight, although I have extremely low confidence in this even compared to the rest of this post. I’d be interested to know whether people whose work on AI safety focuses on limiting the capacity of people to train the AI think that this could be a concern, but I think that any impacts of cheaper electronics from phytomining would likely be so minor that this isn’t a significant downside. This also means that other upsides (more accessibility of smartphones in the Global South) and downsides (maybe more accessibility of social media to teenagers) of cheaper electronics are also very likely insignificant.

Animal Welfare:

Some of the metals that can be phytomined are toxic to many species, including thallium. Arsenic can also be phytomined in theory, although economically it’s not a particularly viable proposition. This includes humans, but humans can and do take precautions to avoid the toxicity associated with these substances. Plants with high concentrations of metals in them are often toxic to animals; I was told some time ago that this is what locoweed is, but that’s not actually true. Phytomining increases the number of these plants around and puts them near humans, thus exposing them to animals. However, everything reported in van der Ent et al. (2021) suggests that animals usually have the sense to avoid these plants if they are toxic. There might also be a need for herbicide or insecticide in some cases, I’m not sure about that. Phytomining also depletes these metals from the soil, which can in the long run be good for letting plants grow there without unhealthy concentrations of metals. Thus, I don’t judge this to be a particularly significant concern, but it’s probably worth keeping an eye out for any negative effects on animal species (or humans) that pop up as a result of phytomining.

The other possible objection to phytomining on animal welfare grounds is the meat-eater problem. Anything that makes human lives better off will let them eat more meat, which means that helping humans kills animals. My sense is that there are a lot of different justifications for how support for helping animals and present-day humans can co-exist, but if you are concerned about the meat-eater problem, phytomining is one of many interventions to no longer look promising.


I recommend phytomining enthusiastically but with some caveats for further research and study by organizations that share goals with the Effective Altruism community. Phytomining promotion shows the potential to be a promising civilization resilience intervention and is worthy of consideration for its impacts on global development and also climate change. Benefits to global health and biodiversity are worth noting but a brief glance does not make them seem strong enough to be good candidates for further investigation in those areas. Natural resource depletion isn’t a traditional cause area of Effective Altruism, but phytomining could also be extremely valuable for people with interests in that area. However, there’s also a very speculative potential for negative impacts due to phytomining in nuclear risk, AI safety and alignment, and animal welfare. 

There were a few points that seem particularly important for further refining our understanding of the promise of phytomining. Firstly, to what extent does phytomining replace conventional mining and to what extent does it add to the metal supply of conventional mining. My expectation, which I believe to be implied in van der Ent et al. (2021), is that it does some of both. Certain elements with extremely high phytomining potential such as thallium could see both the replacement of conventional mining and some additional supply. However, since there’s a much higher demand for nickel, it would be harder for phytomining to eliminate conventional nickel mining. These considerations are likely to vary by element. From the Effective Altruism perspective, this is an important question because the positive impacts in some cause areas rely on one mechanism or the other. The impacts on climate change, natural resource depletion, artificial intelligence, and nuclear risk mostly rely on it adding to the metal supply. The impacts on global health and development as well as biodiversity and some of the climate change impacts mostly rely on it displacing conventional mining. 

Secondly, it’s important to consider what specific metals can be phytomined and how well they match up to the metals important for different cause areas. This is particularly important for civilizational resilience, where specific metals are likely to have a much larger impact, but it also matters for many of the other cause areas. I’ve already done some preliminary investigation on the subject, as reflected in the post, but there’s a lot more room to examine and quantify the phytomining potential of different metals. 

Thirdly, it’s still an open question as to how bad conventional mining is for global health, including global mental health, and development. I lean towards thinking it’s pretty bad, since there seems to be a stated preference against mines in a lot of communities near the mines and I don’t have particularly strong reasons to question it. However, I wasn’t able to find much information on this subject in my fairly quick look through the area, and it particularly seems to lack the quantitative analyses that could help compare it to other global health cause areas. Polls done in the developing world could make significant progress towards answering this question.

Lastly, a number of the impacts, particularly in global health and climate change, are reliant on getting the metal out of the plants being more energy-efficient than traditional smelting. This view is well supported by van der Ent et al. (2021) and Morse (2020), but arguments against it would be bad for the case for phytomining. Alternatively, I haven’t looked much at other ways to produce metal with greater energy-efficiency, but it’s also possible that phytomining could look much worse if interventions with similar but larger effects are also available.

For my particular views about cause area prioritization, this makes phytomining seem quite promising. However, I’d expect that other people could have radically different moral and practical views that could affect their cause area prioritization. I already mentioned the meat-eater problem, but differing views on how bad extinction is compared to catastrophe would also affect the expected value of civilization resilience interventions. Thus, it’s important to note that phytomining being a cause that crosses cause areas is particularly good if you think a lot of cause areas are valuable, but if you think that one cause area is particularly valuable, its breadth does not help it.

I will also venture a quick qualitative Importance, Neglectedness, Tractability analysis given my personal views on phytomining as well as the relative value of different cause areas. I have a lot of uncertainty about importance in particular, but it seems like it could be either a little less or a little more important than other common problems in Effective Altruism. I think specific parts of phytomining, especially work on policy and scaling, are extremely neglected, but botany-heavy work is less neglected. I’d also venture that the tractability is probably above-average but a little odd; it seems that there will eventually be strong economic incentives to do phytomining, but that it carries a lot of uncertainty now. This is influenced by how many open questions in phytomining are specific to certain metals, plants, or regions of the world. A small successful phytomining company could do well to inspire others, but the counterfactual impact of a specific small company is hard to imagine. Something that doesn’t fit neatly into these three categories is that some forms of expertise useful for phytomining (plants, soils, Southeast Asia, mining) aren’t in high demand, as far as I know, in other causes of Effective Altruism.

Next Steps:


The obvious next step, which I’ve called for multiple times in this piece already, is some form of further research and exploration about the potential of phytomining as a cause area of Effective Altruism, likely focusing on its potential to help with one of climate change, global health and development, or civilizational resilience. However, there would be some benefits to keeping the broad-spectrum approach of seeing what phytomining promotion could do that I have adopted in this piece. A broad approach that looks at multiple different mechanisms by which phytomining might improve the world is better at identifying downsides to phytomining, and could consolidate research that would otherwise be multiple different reports. On the other hand, it would be really hard to produce a coherent quantitative estimation of the value of phytomining across cause areas. It also would probably be a challenge to make future recommendations for more concrete actions on phytomining if they don’t strongly target one cause area, because so much funding in Effective Altruism is cause-area-specific.

Based on my experience, I’d advise anybody undertaking a longer research project about phytomining to start with van der Ent et al. (2021); it’s not quite the state of art of the field but is very close. Anything mentioned in Krol-Sinclair and Hale (2023) is probably a good place to start for a deeper dive into policy levers. Furthermore, I’d very strongly advise contacting experts in the field; I didn’t do so both because of time constraints on my part and because I didn’t want to waste their time for something that was supposed to be shorter than it ended up being. These experts could include people working for phytomining startups like Econick and LIFE-AGROMINE, academics such as those at the Center for Mined Land Rehabilitation, and scientists from outside the phytomining community who can provide an outside perspective on phytomining’s claims as a field. 

I’d also recommend this report being written by somebody with more experience in relative fields than I have; even if not a geologist, environmental scientist, or botanist, somebody with any familiarity with engineering and chemistry should be able to produce better estimates of the phytomining potential of specific metals than I have. I also think that, although I’ve mostly eschewed quantitative claims here, quantitative reasoning would be very useful to be able to more explicitly compare phytomining to other causes and interventions.

After Research:

Even by the standards of this post, this section consists of speculative things I think such a further research report could recommend. One possible tool for phytomining promotion would be a charity to work on advocating for pro-phytomining policies, likely starting in Indonesia or maybe Southeast Asia as a whole. It wouldn’t need to be particularly large and I doubt the marginal return would be particularly high from it being more than a few people. 

If civilizational resilience is the main concern, one promising intervention would be to campaign for the Svalbard Global Seed Vault to open its doors to non-crop seeds (Svalbard Global Seed Vault n.d.). Currently the seed vault prioritizes crop plants and their close genetic relatives, but in many possible catastrophes, having both proven medicinal plants and hyperaccumulators would likely be a civilizational priority. I’m honestly not sure why this isn’t already the case; figuring out why it isn’t would likely be important to getting this intervention to work well. Establishing a seed vault just for hyperaccumulators would be possible but costly. Another civilizational-resilience-tinged intervention would be to write a short but practical how-to-phytomine guide and put it with all the other sources of knowledge that people want to survive a major catastrophe.

And, of course, more research! Civilization resilience is hard to run experiments for, because civilization hasn’t yet collapsed (knock on wood), but there are other elements of phytomining, such as health benefits, that seem amenable to high-quality research such as RCTs. 

Other Steps:

I’m not sure if this is correct, but I imagine that Effective Altruism organizations implicitly have a list of things that could affect the sorts of things they care about. I imagine this list to include both activities elsewhere in Effective Altruism-affiliated organizations — for example, what AI regulations might do to non-AI causes — and also other important civilization developments, like cheaper solar power or the Russo-Ukrainian War. I think phytomining belongs on that list. It spans a wide variety of topics that the Effective Altruism community cares about, as demonstrated by how many cause areas I discussed above. It’s easy to imagine a future where phytomining becomes dominant for many metals, but also one where it fades into the dustbin of weird inventions. 

So obviously, if there are explicit lists of these kinds of sources of disruption to consider, I’d urge people to add phytomining to them. Explicitly considering phytomining in forecasts could also be good, if you’re convinced of its potential. But I’ll go further and say that more writing on phytomining would be positive in general. There isn’t a lot of non-technical descriptions out there about phytomining that really go beyond the surface level, although Morse (2020) is pretty good. More information about phytomining that is readily available would help get the word out to potential stakeholders, such as mainstream climate activists and the parts of the global development field that are not part of Effective Altruism.


All sources are cited in Chicago Author-Date format, more or less. Thanks for staying with this piece to the end; any and all feedback is welcome! If you'd rather provide feedback anonymously, you can do so here; let me know if I'm allowed to post my response to it in the comments section.


  1. Alchemia Nova. n.d. “LIFE-AGROMINE.” Accessed July 17, 2023. https://www.alchemia-nova.net/projects/agromine/
  2. Centers for Disease Control and Prevention. n.d. “NIOSH Mining.” Accessed July 18, 2023. https://wwwn.cdc.gov/niosh-mining/MMWC/Fatality/NumberAndRate
  3. Dartnell, Lewis. 2014. The Knowledge: How to Rebuild Our World from Scratch. London: The Bodley Head. https://gizmodo.com/everything-you-need-to-know-to-rebuild-civilization-fro-1566170266
  4. Dattani, Saloni, Fiona Spooner, Hannah Ritchie, and Max Roser. 2023. “Causes of Death.” Our World in Data. https://ourworldindata.org/causes-of-death
  5. Hobson, Peter. 2019. “More than 40 million people work in artisanal mining: report.” Reuters, April 24, 2019. https://www.reuters.com/article/us-mining-asm/more-than-40-million-people-work-in-artisanal-mining-report-idUSKCN1S025C
  6. IEEE. n.d. “Semiconductor Materials.” Accessed July 22, 2023. https://irds.ieee.org/topics/semiconductor-materials
  7. IndustriALL. n.d. “Mining and DGOJP.” Accessed July 18, 2023. https://www.industriall-union.org/mining-and-dgojp
  8. Kidd, Petra Susan, Aida Bani, Emile Benizri, Cristina Gonnelli, Claire Hazotte, Johannes Kisser, and Maria Konstantinou et al. 2018. “Developing Sustainable Agromining Systems in Agricultural Ultramafic Soils for Nickel Recovery.” Frontiers in Environmental Science, June 08, 2018. https://doi.org/10.3389/fenvs.2018.00044
  9. Krol-Sinclair, Maria J. and Thomas Hale. 2023. “The United States Needs to Innovate New Mineral Production Technologies. Here’s One.” Center for Strategic and International Studies, March 24, 2023. https://www.csis.org/analysis/united-states-needs-innovate-new-mineral-production-technologies-heres-one
  10. Malhotra, Madhav. 2022. “Biodiversity Loss: In the Shadow of Climate Change.” Effective Altruism Forum, May 9, 2022. https://forum.effectivealtruism.org/posts/5QTvWMTamPnMmGfyA/biodiversity-loss-in-the-shadow-of-climate-change
  11. Morse, Ian. 2020. “Down on the Farm That Harvests Metal from Plants.” New York Times, February 26, 2020. https://www.nytimes.com/2020/02/26/science/metal-plants-farm.html
  12. Nkrumah, Philip Nti, Romane Tisserand, Rufus L. Chaney, Alan J.M. Baker, Jean Louis Morel, Romain Goudon, Peter D. Erskine, Guillaume Echevarria, and Antony van der Ent. 2018. “The first tropical ‘metal farm’: Some perspectives from field and pot experiments.” Journal of Geochemical Exploration, December 11, 2018. https://doi.org/10.1016/j.gexplo.2018.12.003
  13. Pitchbook. n.d. “Econick Overview.” Accessed July 17, 2023. https://pitchbook.com/profiles/company/277765-30#team
  14. Roberts, David. 2022. “Here are the metals we need for batteries, solar and other clean energy tech.” Canary Media, February 7, 2022. https://www.canarymedia.com/articles/clean-energy/the-minerals-used-by-clean-energy-technologies
  15. Sajid, Haziqa. 2023. “Can You Build Large Language Models Like ChatGPT At Half Cost?” Unite.AI, May 11, 2023. https://www.unite.ai/can-you-build-large-language-models-like-chatgpt-at-half-cost/
  16. Stephens, Carolyn and Mike Ahern. 2001. “Worker and Community Health Impacts Related to Mining Operations Internationally: A Rapid Review of the Literature.” Mining, Minerals, and Sustainable Development 25. https://www.iied.org/sites/default/files/pdfs/migrate/G01051.pdf
  17. Timmerman, Antonia. 2022. “The dirty road to clean energy: How China’s electric vehicle boom is ravaging the environment.” Rest of World, November 28, 2022. https://restofworld.org/2022/indonesia-china-ev-nickel/
  18. United States Environmental Protection Agency. 2012. “A Citizen’s Guide to Phytoremediation.” https://www.epa.gov/sites/default/files/2015-04/documents/a_citizens_guide_to_phytoremediation.pdf
  19. University of Queensland. n.d. “Leaders of the energy transition are calling for a sustainable source of critical metals — is phytomining the answer?” Sustainable Minerals Institute. https://smi.uq.edu.au/leaders-energy-transition-sustainable-source-critical-metals-phytomining
  20. Van der Ent, Antony, Alan J.M. Baker, Roger D. Reeves, Rufus L. Chaney, Christopher W.N. Anderson, John A. Meech, Peter D. Erskine et al. 2015. “Agromining: Farming for Metals in the Future?” Environmental Science and Technology, February 20, 2015. https://doi.org/10.1021/es506031u
  21. Van der Ent, Antony, Alan J.M. Baker, Guillaume Echevarria, Marie-Odile Simonnot, and Jean Louis Morel eds. 2021. Agromining: Farming For Metals. 2nd ed. Cham: Springer Nature Switzerland AG.
  22. Wiblin, Robert. 2016. “Why don’t many effective altruists work on natural resource scarcity?” Effective Altruism Forum, February 20, 2016. https://forum.effectivealtruism.org/posts/2vmiMQjNAH7rQcmz5/why-don-t-many-effective-altruists-work-on-natural-resource.
  23. World Nuclear Association. 2023. “Supply of Uranium.” Last modified May, 2023. https://www.world-nuclear.org/information-library/nuclear-fuel-cycle/uranium-resources/supply-of-uranium.aspx





More posts like this

Sorted by Click to highlight new comments since:

I really appreciate your insights into and estimates on this potential cause area! I have heard of using plants to identify where there are precious minerals, and to concentrate radioactive material but not literal plant mining before now.

Apologies if you covered this in the article, but can phytomining scale? Does it have the potential to be an economical answer and major source of material. Or is it more for special circumstances where mining is unusually difficult for some reason eg the mineral is highly dispersed. If you answered this already please just link me to the section, I was skimming rather haphazardly.

I think I'll go back now and read closer the different types of minerals discussed, but I wanted to write a comment expressing my appreciation before I got distracted. There's not enough novel ideas explored thoughtfully and this is the kind of content that really makes EA shine in my opinion.

I cut out a bit that I had on how much it could scale to make it shorter, but there's still a bit about that in the practical section. In general, phytomining could potentially produce the vast majority of some of the less common elements (thallium), and Krol-Sinclair and Hale say that the same is true for cobalt. It could also produce a pretty large chunk of nickel- I don't have an exact number but I'm pretty confident that it could be a double-digit percentage of the world's consumption. So yes, it scales, but figuring out how much it scales would be a good target for future research.

I was not expecting this to be the answer. That's really fascinating. Phytomining is officially on my radar now, and I'll be linking back to this article. I hope attention to it starts taking off.

Curated and popular this week
Relevant opportunities