Ai Pollution: Is It Real Or Just Hype?

Artificial intelligence (AI) is increasingly being used to tackle some of the world's biggest environmental emergencies, such as tracking climate change and pollution in real time, predicting future causes and areas of pollution, and driving innovations in anti-pollution technology. However, AI also contributes to air, water, and earth pollution in several ways. Firstly, producing AI models requires a lot of energy, which often comes from burning fossil fuels. Secondly, the IT equipment used in AI relies on non-renewable resources that are mined in environmentally damaging ways. Lastly, the disposal of IT equipment at the end of its lifecycle can result in electronic waste that further pollutes the environment. While AI has the potential to address global challenges, its environmental impact and contribution to pollution are concerns that need to be addressed.

AI's energy usage

Data centres, which house most large-scale AI deployments, are large consumers of electricity. AI models, particularly generative AI models, are becoming exponentially larger, and more data centre energy is needed to train them and process data. The production, transport, maintenance, and disposal of hardware components also require substantial energy use and natural resources.

The energy used to power AI often comes from burning fossil fuels, which emit pollutants and are the biggest contributor to global warming and climate change. The generation of electricity from fossil fuels also impacts the environment through the extraction of fossil fuels and energy infrastructure, which can disrupt ecosystems.

However, it is important to note that the efficiency of AI software and hardware in using energy has improved. Some researchers predict that AI's carbon footprint will soon plateau and then shrink due to these improvements. Additionally, AI can be designed to be more energy-efficient, and some tools can help reduce the carbon intensity of AI operations.

The environmental impact of AI is a complex issue, and while it has negative effects, it also has the potential to help tackle some environmental emergencies, such as mapping destructive sand dredging and charting emissions of methane.

AI's electronic waste

AI's environmental impact is a pressing concern. The development, maintenance, and disposal of AI technology all come with a large carbon footprint. AI is powered by massive amounts of energy, much of which comes from the burning of fossil fuels, the biggest contributor to global warming. The electricity generation used to power AI emits pollutants, especially in regions where fossil fuels dominate electricity generation.

The mining and production of metals used in AI hardware can also lead to soil erosion and pollution. The materials used in these devices can contaminate soil and water when not disposed of correctly. The data centres that house AI servers produce electronic waste and are large consumers of water, which is becoming scarce in many places. They also rely on critical minerals and rare elements, which are often mined in environmentally destructive ways.

The rise of AI will increase the scale of e-waste. By 2030, annual e-waste production is expected to reach 75 million metric tons, and by 2050, the World Economic Forum (WEF) projects that the total amount of e-waste generated will have surpassed 120 million metric tons. However, only 17% of global e-waste is documented to be collected and properly recycled each year.

AI-driven e-waste management systems can help address the challenge of e-waste. AI-powered robotic systems can disassemble electronic devices, separating valuable components for recycling. AI improves the recovery and reuse of materials from e-waste, reducing the environmental impact of mining for raw materials. It also contributes to resource conservation by efficiently reclaiming valuable materials like gold, silver, and copper. AI plays a significant role in ensuring the optimal functioning of recycling machinery through predictive maintenance algorithms that minimize downtime and enhance efficiency.

To address the challenges of capacity growth and e-waste, low-to-no-carbon power sources will be key. The focus on renewable energy sources and microgrids will drive innovations in backup power sources such as batteries, hydrogen, and nuclear energy.

AI's impact on consumer behaviour

AI has the potential to positively and negatively impact consumer behaviour concerning pollution and the environment.

AI can be used to reduce climate change and its impact on the environment. AI is being used to map the destructive dredging of sand and chart emissions of methane, a potent greenhouse gas. AI can also be used to enhance energy efficiency and reduce energy usage, as well as assist in environmental monitoring and management, such as tracking air emissions. AI is also being used to flag emissions from global pollution sources. Scientists identify the emissions coming from monitored facilities, and engineers train algorithms on the data to teach the programs to estimate emissions based on visual input alone. These numbers can then be used to inform corporations and policymakers and hold polluters accountable.

However, AI can also be used to negatively impact the environment. AI-powered self-driving cars, for example, could cause more people to drive instead of cycling or taking public transport, pushing up greenhouse gas emissions. AI can also be used to generate misinformation about climate change, downplaying the threat in the public eye. AI is also used in automated advertising, boosting overall consumptive behaviour in society. For instance, with fast-fashion advertising, targeted ads push a steady rotation of cheap, mass-produced clothes to consumers, creating a higher demand for fast-fashion companies. The fashion industry is collectively estimated to produce up to eight per cent of global emissions.

AI's hardware production

AI hardware production has a significant environmental impact. The production, transport, maintenance, and disposal of AI hardware components, such as servers and data centers, require substantial energy use and natural resources. The mining and production of metals used in AI hardware can lead to soil erosion and pollution. The materials used in these devices can contaminate soil and water if not disposed of correctly, contributing to electronic waste.

The manufacturing of hardware, such as GPUs (graphical processing units), requires energy, and the more complex the AI model, the more powerful hardware it needs to run. GPUs are often on servers within data centers, which are large consumers of electricity and water. Data centers are estimated to use between 40% and 50% of their energy for powering computing equipment, with an additional 30% to 40% dedicated to cooling the equipment.

The microchips that power AI need rare earth elements, which are often mined in environmentally destructive ways. This mining process can lead to water usage, soil erosion, and the emission of planet-warming greenhouse gases. The disposal of electronic waste from data centers can also be hazardous, as it often contains toxic substances like mercury and lead.

To mitigate these negative impacts, transitioning to renewable energy sources, adopting energy-efficient practices, and improving recycling and disposal methods for hardware components are essential. Standardized procedures for measuring the environmental impact of AI and regulations for companies to disclose the direct ecological consequences of their AI-based products are also necessary steps.

AI's data centres

AI data centres are a significant source of pollution. They produce electronic waste, consume large amounts of water, and rely on critical minerals and rare elements that are often mined unsustainably.

Firstly, AI data centres produce electronic waste, which often contains hazardous substances such as mercury and lead. The disposal of these hazardous materials can cause soil and water contamination if not done properly.

Secondly, data centres are large consumers of water, which is becoming an increasingly scarce resource. Millions of litres of water are pumped into data centres to keep them cool, and this water usage can strain access to clean water for communities.

Thirdly, data centres rely on critical minerals and rare earth elements, such as cobalt and silicon, which are necessary for hardware devices and microchips. However, the mining of these materials is often done in environmentally destructive ways, leading to soil erosion and pollution.

Furthermore, data centres that host AI technology have massive energy demands, which are typically met by burning fossil fuels. This energy consumption contributes to the emission of planet-warming greenhouse gases, leading to climate change and air pollution. The International Energy Agency estimates that by 2026, electricity consumption by AI data centres could reach 4% of annual global energy usage.

To address these environmental challenges, countries can encourage companies to adopt renewable energy sources and offset carbon emissions in their data centres. Additionally, stronger regulations and transparent reporting can help mitigate health risks as AI adoption grows.

Frequently asked questions