Environmental Impact Insights

The Water Footprint of AI: Why We Need to Pay Attention to Its Environmental Cost As artificial intelligence continues to advance, its environmental impact, particularly concerning water consumption in data centres, warrants attention. Understanding AI's Water Usage AI models, especially large language models, require substantial computational resources. This computing power, concentrated in data centres, generates significant heat, necessitating extensive cooling, often through water-based systems. - Per Query Water Usage: Each interaction with AI models like ChatGPT consumes water. For instance, a 20-50 question session can use approximately 500 millilitres of water, primarily for cooling purposes. - Industry Impact: Data centres globally consumed over 660 billion liters of water in 2022 to cool servers running various services, including AI workloads. Key Areas of Concern 1. Water Scarcity: Many data centres are located in regions with limited water resources. In areas like California, where numerous tech companies operate, water-intensive cooling for AI adds strain to local supplies. 2. Seasonal Impact: During summer, data centres often double their water usage to maintain optimal temperatures. With climate change leading to more frequent heatwaves, this demand could increase, exacerbating the impact. 3. Comparative Impact: Training large AI models can consume up to five times more water than traditional data center operations, highlighting the need for efficient resource management. Steps Toward Sustainability To foster a more sustainable AI ecosystem, the tech industry can consider the following measures: 1. Adopt Alternative Cooling Solutions: Implementing methods like liquid immersion cooling, direct air cooling, and utilising recycled water systems can reduce water demands by up to 90% in certain environments. 2. Enhance Transparency and Accountability: Publicly reporting water usage and environmental impact data allows companies to foster accountability and enable informed consumer choices. Currently, only a few tech giants release detailed sustainability reports on water use. 3. Optimise Model Efficiency: Redesigning models to perform with lower computational intensity can significantly reduce both water and energy requirements. Model efficiency improvements, even by 10-15%, can save millions of litres of water annually. While AI offers transformative benefits across various sectors, it's crucial to balance its growth with responsible resource use. Focusing on sustainable AI practices is essential not only for environmental preservation but also for the technology's long-term viability.By embracing these strategies, we can ensure AI's advancement doesn't come at the expense of our planet's resources. Visual: The Times #ai #waterconsumption #sustainability #datacenters #environmentalimpact #greenai

Can artisanal gold mining be less damaging to the environment? For at least 16 million people worldwide, artisanal small-scale gold mining (ASGM) is a pillar of stability and opportunity, particularly in rural, impoverished communities. But the industry is responsible for a great deal of environmental damage, such as deforestation and contamination: Mining requires the use of harmful chemicals such as mercury, which pollutes air, soil and water, threatening biodiversity and human health. Aimee Gabay reports on efforts to make artisanal small-scale gold mining less damaging, including the U.N. Minamata Convention on Mercury, an international treaty to regulate and eradicate mercury use, which came into force in 2017. Countries such as Ghana, which ratified the agreement in 2017, have laws to regulate the industry and safeguard the environment, but implementation has been weak, according to industry experts. Some ways people are trying to reduce the environmental impact of ASGM include: 🛑 Eliminating mercury use: A key approach is promoting mercury-free alternative technologies, like enhanced gravity concentration, which rely on the natural density of gold to separate it from other materials without using mercury. These alternatives need to be proven advantageous to miners to encourage their adoption. 🏫 Education and support: Providing intensive education to miners about the health and environmental dangers of mercury and offering guidance on using mercury-free techniques is crucial. For instance, in the Philippines, educational efforts coupled with economic factors, like the high price of mercury, have helped miners transition to safer methods. ⚖ Formalizing and regulating ASGM: Governments are encouraged to create and enforce regulations that help formalize ASGM, making it easier for miners to operate legally and responsibly. This includes streamlining licensing processes and ensuring that miners can access formal markets and training. 🌈 Coexistence models: In some areas, artisanal miners collaborate with large-scale mining companies by selling their ore to be processed responsibly, which reduces environmental impacts and helps with formalization. 💰 Addressing socioeconomic factors: To encourage miners to adopt cleaner alternatives, these methods must be economically viable, taking into account the miners' time, costs, and existing practices. 📰 Can nations ever get artisanal gold mining right? https://lnkd.in/gjikGJ4b

Microplastic pollution is disrupting photosynthesis across ecosystems and cropland, threatening our food security in ways we've drastically underestimated. These invisible particles degrade our soils and undermine the biological foundation of our food system. Microplastics physically block sunlight on leaf surfaces, disrupt internal nutrient transport pathways, release adsorbed toxins into plant tissues, and induce oxidative stress—collectively degrading photosynthetic efficiency at cellular levels. Research shows microplastic particles in soil and water reduce plants' ability to convert sunlight to energy by 7-12% and decrease the green pigment plants need for this process by 11-13%. This damage affects everything from crop fields to marine algae. Translating these findings into agricultural yields reveals a shocking reality: we're losing 109-360 million metric tons of crops annually. This massive reduction represents enough food to feed hundreds of millions of people and constitutes 4-14% of global staple crop production. Agriculture exists in a paradoxical relationship with microplastics—both victim and vector. Recent UK soil sampling found agricultural soils contain 1,320-8,190 microplastic particles per kg, with significantly higher concentrations where plastic crop covers were used. Plastics used to increase yields (mulches, covers, polytunnels) degrade into microplastics that subsequently reduce those same yields through photosynthetic interference, creating a classic ecological trap in agricultural systems. To ensure an abundant and healthy food supply, reducing and remediating plastic pollution needs to become a priority in agricultural and environmental policies. Until then, microplastics are undermining the most fundamental biological process that sustains crop productivity.

It is time for #NVIDIA, #AMD, #Intel and every other AI hardware vendor to stop hiding the true environmental cost of their products. Billions are being invested in AI hardware, yet we still lack transparent data on embodied emissions, resource use, water intensity and toxicity impacts. NVIDIA (et al) release selective impact assessments designed to meet compliance needs but conveniently exclude everything that matters. A great new study helps to fill some key gaps: “More than Carbon: Cradle-to-Grave Environmental Impacts of GenAI Training on the NVIDIA A100 GPU”. Its got so much valuable info and is worth a read. https://lnkd.in/eSwzd624 Unlike most studies that rely on secondary data, the researchers physically dismantled an NVIDIA A100 GPU ground it up and carried out a full elemental composition analysis. Using that data, they modelled sixteen environmental impact categories across the entire life cycle, covering raw material extraction, manufacturing, model training and end-of-life. The findings are so interesting: 1. Manufacturing is the dominant source of impact a) Manufacturing a accounts for 81.8% of the total climate impact and 80% of fossil resource depletion before it trains a single model. b) 71% of mineral and metal depletion and 94.5% of cancer-related human toxicity impacts occur during manufacturing. c) The copper-heavy heatsink alone is responsible for 91% of cancer related toxicity, 86% of freshwater eutrophication and 91% of land use impacts. d) Semiconductor fabrication at 7nm is a hotspot, with each square cm of silicon requiring significantly more energy, chemicals and water than previous generations. 2. Training is highly energy intensive but not the whole story a) Training GPT-4 on A100s consumed the equivalent of 11,522 people’s annual climate-change budget. b) In Iowa, where GPT-4 was trained, the carbon-intensive grid drives 96.8% of the training climate footprint. c) Focusing on energy efficiency alone will not solve the problem. Operational carbon dominates the impact, but toxicity, water stress and mineral depletion are driven by manufacturing. 3. AI’s material dependency is huge and invisible a) An A100 contains dozens of rare earths & critical minerals including copper, gold, palladium, platinum and tantalum. b) They found a 33% increase in mineral and metal depletion impacts compared with standard LCAs (i.e. secondary data significantly underestimates things). c) Semiconductor fabrication is concentrated in water-stressed regions such as Taiwan, South Korea and Arizona, yet vendors do not disclose water intensity per GPU. This is why hardware vendors must conduct full component level PCF's incl. verifiable embodied impact data. Without transparency we are literally flying blind whilst they make trillions of dollars. This study is an important milestone, but it also shows how little we really know about the environmental impact of AI hardware. Vendors... stop hiding

Research has highlighted the environmental impact of generative AI, particularly as it relates to the energy demands of data centers. A recent Morgan Stanley report predicts that AI-related industries could emit up to 2.5 billion tons of greenhouse gases by 2030, largely due to the growing need for data centers to support AI workloads. The Green Software Foundation(GSF) Software Carbon Intensity (SCI) Specification provides a practical framework for addressing these concerns. While SCI is applicable to all software, its core principles are particularly impactful in reducing the carbon footprint of AI systems, with the goal being to reduce emissions actively, not just offset them: 1️⃣ Energy Efficiency: Optimizing AI models to use less energy is critical. Techniques like model pruning and distillation help make AI models more efficient by reducing the number of parameters and complexity without sacrificing performance, thus cutting down the energy required for training and deployment. 2️⃣ Hardware Efficiency: Using energy-efficient chipsets and maximizing hardware utilization can help reduce emissions from AI workloads. This involves developing hardware that can handle AI computations more efficiently and extending the lifecycle of existing hardware to reduce the need for frequent replacements, which contribute to emissions during production and disposal. 3️⃣ Carbon Awareness: AI systems can be made carbon-aware, meaning workloads are scheduled to run when energy grids are powered by cleaner, renewable energy. This minimizes the reliance on carbon-intensive power sources and reduces the overall environmental impact. For meaningful progress, policymakers must implement robust regulatory frameworks that support these efforts. Regulations that enforce carbon reporting for AI systems, incentivize the use of renewable energy, and establish standards for emissions will be key to aligning the AI industry with global sustainability goals. By integrating SCI principles with strong policy support, the AI industry can make substantial strides in reducing emissions while continuing to innovate responsibly. (Link - https://lnkd.in/drMQhDEY) #greenai #sustainability #genai

We should have known this was coming. "Hoof it through the national parks of the western United States—Joshua Tree, the Grand Canyon, Bryce Canyon—and breathe deep the pristine air. These are unspoiled lands, collectively a great American conservation story. Yet an invisible menace is actually blowing through the air and falling via raindrops: Microplastic particles, tiny chunks (by definition, less than 5 millimeters long) of fragmented plastic bottles and microfibers that fray from clothes, all pollutants that get caught up in Earth’s atmospheric systems and deposited in the wilderness. Writing in the journal Science, researchers report a startling discovery: After collecting rainwater and air samples for 14 months, they calculated that over 1,000 metric tons of microplastic particles fall into 11 protected areas in the western US each year. That’s the equivalent of over 120 million plastic water bottles. “We just did that for the area of protected areas in the West, which is only 6 percent of the total US area,” says lead author Janice Brahney, an environmental scientist at Utah State University. “The number was just so large, it's shocking.” It further confirms an increasingly hellish scenario: Microplastics are blowing all over the world, landing in supposedly pure habitats, like the Arctic and the remote French Pyrenees. They’re flowing into the oceans via wastewater and tainting deep-sea ecosystems, and they’re even ejecting out of the water and blowing onto land in sea breezes. And now in the American West, and presumably across the rest of the world given that these are fundamental atmospheric processes, they are falling in the form of plastic rain—the new acid rain."

🌍 Plastics pollution: An underestimated crisis with planetary consequences 🌍 A new study by researchers from the Stockholm Resilience Centre (SRC) uncovers the profound and systemic impacts of plastics pollution on our planet. Published in One Earth, the study reveals that plastics pollution breaches planetary boundaries and amplifies the impacts of others, such as climate change and biodiversity loss. For too long, plastics have been seen as a standalone issue—mainly a waste problem. This study makes it clear: plastics pollution is a systemic, lifecycle challenge that demands urgent action. 🔍 Key takeaways: 1️⃣ Cross-boundary impacts: Beyond the direct harm to ecosystems, plastics pollution intensifies the effects of greenhouse gas emissions, disrupts nutrient cycles, and degrades critical water systems. 2️⃣ Global footprint: Plastics are now ubiquitous—from the ocean floor to polar regions—making their impacts pervasive and hard to manage. 3️⃣ Legacy pollution: Plastics produced decades ago continue to affect ecosystems, while current production rates (504 million tons in 2022) remain dangerously unsustainable. 4️⃣ Environmental justice matters: The study also highlights the inequities of plastics pollution, such as “waste colonialism,” where the burdens disproportionately affect vulnerable communities. 5️⃣ A systems approach is needed: Addressing plastics pollution requires lifecycle-based interventions, global governance, and stronger policies targeting production and chemical transparency. The Stockholm Resilience Centre (SRC) underscores the importance of rethinking how we view plastics—not as inert materials but as complex pollutants that harm the environment and human health throughout their lifecycle. 💡 How is your organization addressing plastics pollution? Have you evaluated your supply chains, procurement, or waste practices to reduce reliance on plastics or improve lifecycle management? Read the full study here: https://lnkd.in/exZe9XDj Related article here: https://lnkd.in/e5_JXS8v #Sustainability #PlasticsPollution #PlanetaryBoundaries #EnvironmentalJustice #ResilienceThinking

We often talk about how revolutionary AI is—transforming industries, boosting productivity, and redefining creativity. But there’s a side to this technology we rarely discuss: its hidden environmental cost. 🌍💧 Did you know that every AI prompt—even the ones we use to write emails or brainstorm ideas—requires immense energy and water? Data centers powering these systems consume vast resources, contributing to water scarcity, increased carbon emissions, and environmental stress. With tools like ChatGPT, DALL-E, and others becoming a daily part of our lives, the scale of this impact is growing faster than we can imagine. In my latest video, I dive into: 👉 The environmental toll of AI, including how it affects water and energy usage. 👉 The connection between tech-driven resource depletion and real-world crises like wildfires. 👉 What big tech companies are doing to combat the issue. 👉 Simple steps we can take as consumers to use AI responsibly. This isn’t about shaming innovation—it’s about responsibility. We must balance advancing technology and preserving the planet for future generations. 🌱 Let me know your thoughts about AI and its environmental impact!

🌍 AI Expansion and Sustainability: The True Cost of Data Centers The era of artificial intelligence has brought transformative advancements across industries, but it has also raised significant concerns about the environmental and social impacts of rapid data center growth. Here are three key points to consider about the challenges we face: 1) Water and Energy Consumption: A study by The Washington Post and the University of California, Riverside revealed that chatbots like GPT-4 can consume up to half a liter of water per email generated. Imagine the impact if one in ten U.S. workers used this technology twice a week: annual water consumption could reach 870 million liters, equivalent to the domestic water use of Rhode Island, on the U.S. East Coast, for three days. 2) Focus on Environmental Justice: The “Third Wave” of AI ethics, discussed by experts like Aimee van Wynsberghe, urges us to consider not just transparency and privacy, but also the socio-environmental impacts on vulnerable communities. Cases in Memphis, U.S., and Santiago, Chile, illustrate how regions with less decision-making power bear the environmental consequences of data centers, worsening social inequalities. 3) Transparency and Sustainability: The use of Renewable Energy Certificates (RECs) allows companies to claim carbon neutrality without actually using clean energy. This underscores the urgent need for stronger policies that demand genuine environmental responsibility, not just surface-level claims. These insights highlight the pressing need for AI governance that balances innovation with environmental and social responsibility. Want to dive deeper into how these issues are shaping our future? 🌱✨ 🔗 Check out the full post and share your thoughts! #SustainableAI #AIethics #DataCenters #EnvironmentalJustice #AIGovernance

Is AI's Growth Sustainable? How to Make Generative Applications Greener. The rise of generative AI tools like ChatGPT and others has been remarkable, but their environmental impact is often overlooked. The data center industry, housing these systems, accounts for up to 3% of global greenhouse gas emissions, with energy consumption doubling every two years. Hyperscale cloud providers like Amazon AWS, Google Cloud, and Microsoft Azure play a significant role in powering these models, leading to major carbon footprints. Understanding the carbon footprint lifecycle of AI models is crucial. Large generative models consume extensive energy during training, and fine-tuning can be a more energy-efficient option. Inference sessions, though less energy-intensive, involve many more sessions, contributing to ongoing energy consumption. Efforts to reduce energy usage include employing less computationally expensive approaches like TinyML and using large models only when significantly valuable. To make AI greener, companies can use existing models from providers instead of creating new ones. Fine-tuning existing models on specific content domains consumes less energy and provides more value. Utilizing energy sources from carbon-friendly regions and monitoring carbon emissions can significantly reduce AI's environmental impact. Reusing models and resources, incorporating AI activity into carbon monitoring, and encouraging green AI practices are crucial steps in promoting sustainability. 1. Prioritize Fine-Tuning: Instead of training new generative models from scratch, focus on fine-tuning existing models for specific content domains. Fine-tuning consumes less energy and provides more value to businesses. 2. Explore Energy-Conserving Methods: Adopt energy-conserving computational approaches like TinyML for processing data. TinyML allows running ML models on low-powered edge devices, significantly reducing energy consumption. 3. Re-use and Open Source Models: Opt for reusing open-source models instead of creating new ones. Recycling tech can lower the carbon impact of AI practices and reduce the need for energy-intensive model development. 4. Monitor Carbon Emissions: Include AI activity in carbon monitoring practices to understand the carbon footprint of AI-related operations. Share footprint numbers to make informed decisions about AI partnerships. 5. Choose Green Energy Sources: Select cloud providers and data centers that prioritize environmentally friendly power resources. Running AI models in regions with carbon-free energy sources can significantly reduce operational emissions. Have you already considered the impact of using compute-heavy applications on our planet? Are you tracking the impact of compute in your sustainability report? #genai #aivalue #sustainableai #sustainability