In an era of technological breakthroughs, the potential of artificial intelligence (AI) and other advancements to transform industries and societies is unprecedented. From automating processes to revolutionizing healthcare, technology has undoubtedly brought immense benefits. Insilico Medicine used generative AI (Gen AI) to discover a COVID-19 therapeutic drug. ChatGPT, DALL-E, and Stable Diffusion have captured the public’s attention to quick problem-solving and overcoming creative blocks, becoming one of the fastest-growing consumer applications in history. However, this progress comes at a cost often unnoticed: the toll on Earth’s sustainability.
This blog delves into the environmental impact of technological advancements, focusing on AI and other emerging technologies and the actionable strategies for investors like you to ensure ethical progress that aligns with a greener future.
The Environmental Toll of AI
Technological advancements, particularly in AI, have transformed the global landscape. Yet, beneath their promise lies a substantial carbon footprint.
Energy Consumption: AI-powered systems and data centers are notoriously energy-intensive. Training a single AI model can consume as much electricity as a small town, leading to significant carbon emissions. For instance, a study revealed that training large language models like GPT-3 could emit as much CO2 as 125 round-trip flights between New York and Beijing. As we all know, excess amounts of carbon dioxide will increase the global temperature. It all connects back to technology’s overconsumption of energy.
Resource Depletion: The production of AI-driven technologies depends on rare earth elements such as cobalt and lithium, essential for batteries and hardware. Cobalt is used in GPUs and powering data centers, and lithium-powered batteries are the backbone for phones, laptops, and IoT (Internet of Things) devices that run AI models. Other rare earth elements, such as neodymium and dysprosium, are used in magnets for cooling systems and hard drives, which is critical for AI servers.
Estimates suggest that by 2050, demand for lithium could increase by 18 to 20 times and cobalt by 17 to 19 times compared to current levels. The mining processes for these minerals often lead to severe environmental degradation. For instance, lithium extraction from salt flats can deplete local water resources, causing droughts and affecting agriculture. Similarly, unregulated cobalt mining practices in the Democratic Republic of Congo have led to deforestation, soil erosion, and water pollution. Developing new mining projects can take over 16 years, from discovery to production. This delay makes it challenging to respond quickly to rising demands, potentially leading to shortages.
Water Usage: Cooling data centers, a critical component of AI infrastructure, requires vast amounts of water. A mid-sized data center can consume around 300,000 gallons of water daily, while larger facilities may use up to 5 million gallons daily. Google's hyperscale data centers averaged approximately 550,000 gallons per day in 2023, equivalent to the daily water needs of thousands of people. Areas like Arizona and Texas in the United States are experiencing heightened water challenges as they are also home to many new data centers. Approximately 20% of these centers draw from watersheds already under significant stress. Such a massive demand adds to the impending pressure of the predicted water shortage in the U.S.
Why Should This Concern You, The Investor?
The biggest reason you or anybody should be taking this matter seriously is because it concerns our planet, and it is the only home we have. We cannot afford to be ignorant until we find another habitable planet and develop interplanetary transport. So it is our social duty as inhabitants of Earth to protect it and nurture it.
There are also a few business reasons that advocate sustainable practices in corporations. Firstly, a lack of vision towards sustainable practices will impact you where it hurts the most – the returns. Companies with high ESG (Environmental, Social, and Governance) scores have been found to generate 2.6 times higher total shareholder returns. How? Adopting ESG practices means cutting down excess energy consumption and waste and optimizing resources, ultimately saving costs. ESG-focused companies are also better equipped to manage risks associated with regulatory compliance and reputation management, which means almost no penalties or public scrutiny.
Studies prove that consumers prefer companies that are vigilant about their ESG performance. A survey by PDI Technologies reveals that 80% of consumers will pay more for sustainably produced or sourced goods. PwC says people are prepared to spend 9.7% more on products that meet specific environmental criteria, such as being locally sourced or made from recycled materials. A study by NielsenIQ says 78% of U.S. consumers indicated that a sustainable lifestyle is important to them. When consumers are driven by ESG practices, why shouldn’t you pivot to eco-friendly operations?
Pro-environmental practices are the concerns of several international organizations and government bodies, and they are asking companies to align with their carbon-free vision. Sadly, only 16% of the world's largest companies are on track to meet their net-zero emissions goals by 2050, as per Accenture. The same study found that nearly half (45%) of these companies have reported increased carbon emissions instead of reductions. As an investor, you must be concerned about the pressure from such organizations as they impact company reputation and, ultimately, the stock prices. If the government is compelled to introduce strict measures for monitoring and limiting energy use by corporations, it can affect your company’s performance and business goals, too.
So What To Do?
As an investor, you hold the power to shape the future of sustainable technology with your expertise and influence. Here are a few things that should be done immediately:
- Minimize computational cost with optimized algorithms
Motivate entrepreneurs to develop or adopt an accurate energy estimation approach that helps reduce the time and energy needed for training and deploying AI. Several reports have found that energy-efficient approaches don’t always need significant compromises on the quality of AI models. Researchers from Google and the University of California, Berkeley, have shown that the carbon footprint of large language models can be reduced by 100 to 1,000 times with the appropriate choice of algorithms, customized hardware, and energy-efficient cloud data centers.
- Invest in R&D for smaller AI models
Ask the engineering team to reduce the complexity of the AI model to reduce energy consumption and carbon footprint. One experiment by Accenture shows that training a much smaller “student” model, just 6% in size of the original “teacher” model, achieved the same level of accuracy (99%) but consumed 2.7 times less energy. So, you need to invest in more research and development to optimize AI models and reduce their energy consumption.
- Actively participate in policy building
Governments and other environmental organizations can use your help to understand a company’s perspective on energy requirements. This is your opportunity to be a bridge between policymakers and entrepreneurs to develop regulations that bring measurable change in energy use.
The march of technological progress is inevitable, but it must not come at the expense of our planet. AI and other advancements hold immense promise, but their environmental costs cannot be ignored. As we look to the future, the choice is clear: innovate responsibly or risk irreversible damage to our planet. The time to act is now.
References
https://pmc.ncbi.nlm.nih.gov/articles/PMC11396088/
https://www.accenture.com/us-en/blogs/consulting/making-generative-ai-green