Massive benefits of AI come with environmental and human costs. Can AI itself be part of the solution?
Generative AI has a large carbon footprint and other drawbacks. But AI can help mitigate its own harms—by plowing through mountains of data on extreme weather and human displacement.
The recent explosion of generative artificial intelligence tools like ChatGPT and Dall-E enabled anyone with internet access to harness AI’s power for enhanced productivity, creativity, and problem-solving. With their ever-improving capabilities and expanding user base, these tools proved useful across disciplines, from the creative to the scientific.
But beneath the technological wonders of human-like conversation and creative expression lies a dirty secret—an alarming environmental and human cost. AI has an immense carbon footprint. Systems like ChatGPT take months to train in high-powered data centers, which demand huge amounts of electricity, much of which is still generated with fossil fuels, as well as water for cooling. “One of the reasons why Open AI needs investments [to the tune of] $10 billion from Microsoft is because they need to pay for all of that computation,” says Kentaro Toyama, a computer scientist at the University of Michigan. There’s also an ecological toll from mining rare minerals required for hardware and infrastructure. This environmental exploitation pollutes land, triggers natural disasters and causes large-scale human displacement. Finally, for data labeling needed to train and correct AI algorithms, the Big Data industry employs cheap and exploitative labor, often from the Global South.
Generative AI tools are based on large language models (LLMs), with most well-known being various versions of GPT. LLMs can perform natural language processing, including translating, summarizing and answering questions. They use artificial neural networks, called deep learning or machine learning. Inspired by the human brain, neural networks are made of millions of artificial neurons. “The basic principles of neural networks were known even in the 1950s and 1960s,” Toyama says, “but it’s only now, with the tremendous amount of compute power that we have, as well as huge amounts of data, that it’s become possible to train generative AI models.”
Though there aren’t any official figures about the power consumption or emissions from data centers, experts estimate that they use one percent of global electricity—more than entire countries.
In recent months, much attention has gone to the transformative benefits of these technologies. But it’s important to consider that these remarkable advances may come at a price.
AI’s carbon footprint
In their latest annual report, 2023 Landscape: Confronting Tech Power, the AI Now Institute, an independent policy research entity focusing on the concentration of power in the tech industry, says: “The constant push for scale in artificial intelligence has led Big Tech firms to develop hugely energy-intensive computational models that optimize for ‘accuracy’—through increasingly large datasets and computationally intensive model training—over more efficient and sustainable alternatives.”
Though there aren’t any official figures about the power consumption or emissions from data centers, experts estimate that they use one percent of global electricity—more than entire countries. In 2019, Emma Strubell, then a graduate researcher at the University of Massachusetts Amherst, estimated that training a single LLM resulted in over 280,000 kg in CO2 emissions—an equivalent of driving almost 1.2 million km in a gas-powered car. A couple of years later, David Patterson, a computer scientist from the University of California Berkeley, and colleagues, estimated GPT-3’s carbon footprint at over 550,000 kg of CO2 In 2022, the tech company Hugging Face, estimated the carbon footprint of its own language model, BLOOM, as 25,000 kg in CO2 emissions. (BLOOM’s footprint is lower because Hugging Face uses renewable energy, but it doubled when other life-cycle processes like hardware manufacturing and use were added.)
Luckily, despite the growing size and numbers of data centers, their increasing energy demands and emissions have not kept pace proportionately—thanks to renewable energy sources and energy-efficient hardware.
But emissions don’t tell the full story.
AI’s hidden human cost
“If historical colonialism annexed territories, their resources, and the bodies that worked on them, data colonialism’s power grab is both simpler and deeper: the capture and control of human life itself through appropriating the data that can be extracted from it for profit.” So write Nick Couldry and Ulises Mejias, authors of the book The Costs of Connection.
The energy requirements, hardware manufacture and the cheap human labor behind AI systems disproportionately affect marginalized communities.
Technologies we use daily inexorably gather our data. “Human experience, potentially every layer and aspect of it, is becoming the target of profitable extraction,” Couldry and Meijas say. This feeds data capitalism, the economic model built on the extraction and commodification of data. While we are being dispossessed of our data, Big Tech commodifies it for their own benefit. This results in consolidation of power structures that reinforce existing race, gender, class and other inequalities.
“The political economy around tech and tech companies, and the development in advances in AI contribute to massive displacement and pollution, and significantly changes the built environment,” says technologist and activist Yeshi Milner, who founded Data For Black Lives (D4BL) to create measurable change in Black people’s lives using data. The energy requirements, hardware manufacture and the cheap human labor behind AI systems disproportionately affect marginalized communities.
AI’s recent explosive growth spiked the demand for manual, behind-the-scenes tasks, creating an industry described by Mary Gray and Siddharth Suri as “ghost work” in their book. This invisible human workforce that lies behind the “magic” of AI, is overworked and underpaid, and very often based in the Global South. For example, workers in Kenya who made less than $2 an hour, were the behind the mechanism that trained ChatGPT to properly talk about violence, hate speech and sexual abuse. And, according to an article in Analytics India Magazine, in some cases these workers may not have been paid at all, a case for wage theft. An exposé by the Washington Post describes “digital sweatshops” in the Philippines, where thousands of workers experience low wages, delays in payment, and wage theft by Remotasks, a platform owned by Scale AI, a $7 billion dollar American startup. Rights groups and labor researchers have flagged Scale AI as one company that flouts basic labor standards for workers abroad.
It is possible to draw a parallel with chattel slavery—the most significant economic event that continues to shape the modern world—to see the business structures that allow for the massive exploitation of people, Milner says. Back then, people got chocolate, sugar, cotton; today, they get generative AI tools. “What’s invisible through distance—because [tech companies] also control what we see—is the massive exploitation,” Milner says.
“At Data for Black Lives, we are less concerned with whether AI will become human…[W]e’re more concerned with the growing power of AI to decide who’s human and who’s not,” Milner says. As a decision-making force, AI becomes a “justifying factor for policies, practices, rules that not just reinforce, but are currently turning the clock back generations years on people’s civil and human rights.”
Ironically, AI plays an important role in mitigating its own harms—by plowing through mountains of data about weather changes, extreme weather events and human displacement.
Nuria Oliver, a computer scientist, and co-founder and vice-president of the European Laboratory of Learning and Intelligent Systems (ELLIS), says that instead of focusing on the hypothetical existential risks of today’s AI, we should talk about its real, tangible risks.
“Because AI is a transverse discipline that you can apply to any field [from education, journalism, medicine, to transportation and energy], it has a transformative power…and an exponential impact,” she says.
AI's accountability
“At the core of what we were arguing about data capitalism [is] a call to action to abolish Big Data,” says Milner. “Not to abolish data itself, but the power structures that concentrate [its] power in the hands of very few actors.”
A comprehensive AI Act currently negotiated in the European Parliament aims to rein Big Tech in. It plans to introduce a rating of AI tools based on the harms caused to humans, while being as technology-neutral as possible. That sets standards for safe, transparent, traceable, non-discriminatory, and environmentally friendly AI systems, overseen by people, not automation. The regulations also ask for transparency in the content used to train generative AIs, particularly with copyrighted data, and also disclosing that the content is AI-generated. “This European regulation is setting the example for other regions and countries in the world,” Oliver says. But, she adds, such transparencies are hard to achieve.
Google, for example, recently updated its privacy policy to say that anything on the public internet will be used as training data. “Obviously, technology companies have to respond to their economic interests, so their decisions are not necessarily going to be the best for society and for the environment,” Oliver says. “And that’s why we need strong research institutions and civil society institutions to push for actions.” ELLIS also advocates for data centers to be built in locations where the energy can be produced sustainably.
Ironically, AI plays an important role in mitigating its own harms—by plowing through mountains of data about weather changes, extreme weather events and human displacement. “The only way to make sense of this data is using machine learning methods,” Oliver says.
Milner believes that the best way to expose AI-caused systemic inequalities is through people's stories. “In these last five years, so much of our work [at D4BL] has been creating new datasets, new data tools, bringing the data to life. To show the harms but also to continue to reclaim it as a tool for social change and for political change.” This change, she adds, will depend on whose hands it is in.
A company has slashed the cost of assessing a person's genome to just $100. With lower costs - and as other genetic tools mature and evolve - a wave of new therapies could be coming in the near future.
In May 2022, Californian biotech Ultima Genomics announced that its UG 100 platform was capable of sequencing an entire human genome for just $100, a landmark moment in the history of the field. The announcement was particularly remarkable because few had previously heard of the company, a relative unknown in an industry long dominated by global giant Illumina which controls about 80 percent of the world’s sequencing market.
Ultima’s secret was to completely revamp many technical aspects of the way Illumina have traditionally deciphered DNA. The process usually involves first splitting the double helix DNA structure into single strands, then breaking these strands into short fragments which are laid out on a glass surface called a flow cell. When this flow cell is loaded into the sequencing machine, color-coded tags are attached to each individual base letter. A laser scans the bases individually while a camera simultaneously records the color associated with them, a process which is repeated until every single fragment has been sequenced.
Instead, Ultima has found a series of shortcuts to slash the cost and boost efficiency. “Ultima Genomics has developed a fundamentally new sequencing architecture designed to scale beyond conventional approaches,” says Josh Lauer, Ultima’s chief commercial officer.
This ‘new architecture’ is a series of subtle but highly impactful tweaks to the sequencing process ranging from replacing the costly flow cell with a silicon wafer which is both cheaper and allows more DNA to be read at once, to utilizing machine learning to convert optical data into usable information.
To put $100 genome in perspective, back in 2012 the cost of sequencing a single genome was around $10,000, a price tag which dropped to $1,000 a few years later. Before Ultima’s announcement, the cost of sequencing an individual genome was around $600.
Several studies have found that nearly 12 percent of healthy people who have their genome sequenced, then discover they have a variant pointing to a heightened risk of developing a disease that can be monitored, treated or prevented.
While Ultima’s new machine is not widely available yet, Illumina’s response has been rapid. Last month the company unveiled the NovaSeq X series, which it describes as its fastest most cost-efficient sequencing platform yet, capable of sequencing genomes at $200, with further price cuts likely to follow.
But what will the rapidly tumbling cost of sequencing actually mean for medicine? “Well to start with, obviously it’s going to mean more people getting their genome sequenced,” says Michael Snyder, professor of genetics at Stanford University. “It'll be a lot more accessible to people.”
At the moment sequencing is mainly limited to certain cancer patients where it is used to inform treatment options, and individuals with undiagnosed illnesses. In the past, initiatives such as SeqFirst have attempted further widen access to genome sequencing based on growing amounts of research illustrating the potential benefits of the technology in healthcare. Several studies have found that nearly 12 percent of healthy people who have their genome sequenced, then discover they have a variant pointing to a heightened risk of developing a disease that can be monitored, treated or prevented.
“While whole genome sequencing is not yet widely used in the U.S., it has started to come into pediatric critical care settings such as newborn intensive care units,” says Professor Michael Bamshad, who heads the genetic medicine division in the University of Washington’s pediatrics department. “It is also being used more often in outpatient clinical genetics services, particularly when conventional testing fails to identify explanatory variants.”
But the cost of sequencing itself is only one part of the price tag. The subsequent clinical interpretation and genetic counselling services often come to several thousand dollars, a cost which insurers are not always willing to pay.
As a result, while Bamshad and others hope that the arrival of the $100 genome will create new opportunities to use genetic testing in innovative ways, the most immediate benefits are likely to come in the realm of research.
Bigger Data
There are numerous ways in which cheaper sequencing is likely to advance scientific research, for example the ability to collect data on much larger patient groups. This will be a major boon to scientists working on complex heterogeneous diseases such as schizophrenia or depression where there are many genes involved which all exert subtle effects, as well as substantial variance across the patient population. Bigger studies could help scientists identify subgroups of patients where the disease appears to be driven by similar gene variants, who can then be more precisely targeted with specific drugs.
If insurers can figure out the economics, Snyder even foresees a future where at a certain age, all of us can qualify for annual sequencing of our blood cells to search for early signs of cancer or the potential onset of other diseases like type 2 diabetes.
David Curtis, a genetics professor at University College London, says that scientists studying these illnesses have previously been forced to rely on genome-wide association studies which are limited because they only identify common gene variants. “We might see a significant increase in the number of large association studies using sequence data,” he says. “It would be far preferable to use this because it provides information about rare, potentially functional variants.”
Cheaper sequencing will also aid researchers working on diseases which have traditionally been underfunded. Bamshad cites cystic fibrosis, a condition which affects around 40,000 children and adults in the U.S., as one particularly pertinent example.
“Funds for gene discovery for rare diseases are very limited,” he says. “We’re one of three sites that did whole genome sequencing on 5,500 people with cystic fibrosis, but our statistical power is limited. A $100 genome would make it much more feasible to sequence everyone in the U.S. with cystic fibrosis and make it more likely that we discover novel risk factors and pathways influencing clinical outcomes.”
For progressive diseases that are more common like cancer and type 2 diabetes, as well as neurodegenerative conditions like multiple sclerosis and ALS, geneticists will be able to go even further and afford to sequence individual tumor cells or neurons at different time points. This will enable them to analyze how individual DNA modifications like methylation, change as the disease develops.
In the case of cancer, this could help scientists understand how tumors evolve to evade treatments. Within in a clinical setting, the ability to sequence not just one, but many different cells across a patient’s tumor could point to the combination of treatments which offer the best chance of eradicating the entire cancer.
“What happens at the moment with a solid tumor is you treat with one drug, and maybe 80 percent of that tumor is susceptible to that drug,” says Neil Ward, vice president and general manager in the EMEA region for genomics company PacBio. “But the other 20 percent of the tumor has already got mutations that make it resistant, which is probably why a lot of modern therapies extend life for sadly only a matter of months rather than curing, because they treat a big percentage of the tumor, but not the whole thing. So going forwards, I think that we will see genomics play a huge role in cancer treatments, through using multiple modalities to treat someone's cancer.”
If insurers can figure out the economics, Snyder even foresees a future where at a certain age, all of us can qualify for annual sequencing of our blood cells to search for early signs of cancer or the potential onset of other diseases like type 2 diabetes.
“There are companies already working on looking for cancer signatures in methylated DNA,” he says. “If it was determined that you had early stage cancer, pre-symptomatically, that could then be validated with targeted MRI, followed by surgery or chemotherapy. It makes a big difference catching cancer early. If there were signs of type 2 diabetes, you could start taking steps to mitigate your glucose rise, and possibly prevent it or at least delay the onset.”
This would already revolutionize the way we seek to prevent a whole range of illnesses, but others feel that the $100 genome could also usher in even more powerful and controversial preventative medicine schemes.
Newborn screening
In the eyes of Kári Stefánsson, the Icelandic neurologist who been a visionary for so many advances in the field of human genetics over the last 25 years, the falling cost of sequencing means it will be feasible to sequence the genomes of every baby born.
“We have recently done an analysis of genomes in Iceland and the UK Biobank, and in 4 percent of people you find mutations that lead to serious disease, that can be prevented or dealt with,” says Stefansson, CEO of deCODE genetics, a subsidiary of the pharmaceutical company Amgen. “This could transform our healthcare systems.”
As well as identifying newborns with rare diseases, this kind of genomic information could be used to compute a person’s risk score for developing chronic illnesses later in life. If for example, they have a higher than average risk of colon or breast cancer, they could be pre-emptively scheduled for annual colonoscopies or mammograms as soon as they hit adulthood.
To a limited extent, this is already happening. In the UK, Genomics England has launched the Newborn Genomes Programme, which plans to undertake whole-genome sequencing of up to 200,000 newborn babies, with the aim of enabling the early identification of rare genetic diseases.
"I have not had my own genome sequenced and I would not have wanted my parents to have agreed to this," Curtis says. "I don’t see that sequencing children for the sake of some vague, ill-defined benefits could ever be justifiable.”
However, some scientists feel that it is tricky to justify sequencing the genomes of apparently healthy babies, given the data privacy issues involved. They point out that we still know too little about the links which can be drawn between genetic information at birth, and risk of chronic illness later in life.
“I think there are very difficult ethical issues involved in sequencing children if there are no clear and immediate clinical benefits,” says Curtis. “They cannot consent to this process. I have not had my own genome sequenced and I would not have wanted my parents to have agreed to this. I don’t see that sequencing children for the sake of some vague, ill-defined benefits could ever be justifiable.”
Curtis points out that there are many inherent risks about this data being available. It may fall into the hands of insurance companies, and it could even be used by governments for surveillance purposes.
“Genetic sequence data is very useful indeed for forensic purposes. Its full potential has yet to be realized but identifying rare variants could provide a quick and easy way to find relatives of a perpetrator,” he says. “If large numbers of people had been sequenced in a healthcare system then it could be difficult for a future government to resist the temptation to use this as a resource to investigate serious crimes.”
While sequencing becoming more widely available will present difficult ethical and moral challenges, it will offer many benefits for society as a whole. Cheaper sequencing will help boost the diversity of genomic datasets which have traditionally been skewed towards individuals of white, European descent, meaning that much of the actionable medical information which has come out of these studies is not relevant to people of other ethnicities.
Ward predicts that in the coming years, the growing amount of genetic information will ultimately change the outcomes for many with rare, previously incurable illnesses.
“If you're the parent of a child that has a susceptible or a suspected rare genetic disease, their genome will get sequenced, and while sadly that doesn’t always lead to treatments, it’s building up a knowledge base so companies can spring up and target that niche of a disease,” he says. “As a result there’s a whole tidal wave of new therapies that are going to come to market over the next five years, as the genetic tools we have, mature and evolve.”
In this week's Friday Five, an old diabetes drug finds an exciting new purpose. Plus, how to make the cities of the future less toxic, making old mice younger with EVs, a new reason for mysterious stillbirths - and much more.
The Friday Five covers five stories in research that you may have missed this week. There are plenty of controversies and troubling ethical issues in science – and we get into many of them in our online magazine – but this news roundup focuses on scientific creativity and progress to give you a therapeutic dose of inspiration headed into the weekend.
Here is the promising research covered in this week's Friday Five:
Listen on Apple | Listen on Spotify | Listen on Stitcher | Listen on Amazon | Listen on Google
- How to make cities of the future less noisy
- An old diabetes drug could have a new purpose: treating an irregular heartbeat
- A new reason for mysterious stillbirths
- Making old mice younger with EVs
- No pain - or mucus - no gain
And an honorable mention this week: How treatments for depression can change the structure of the brain