What Is AI’s Environmental Impact? Solutions for the Future With AI

If it seems like artificial intelligence (AI) is everywhere these days, that’s because it is. This rapidly developing form of technology can shape every aspect of our lives, and the environment is no exception. But it’s not all benefits. Many have expressed concerns about AI’s environmental impact and potential damage when it comes to energy use, pollution, deforestation and other environmental harms.

Will the sustainability opportunities AI affords offset its potential environmental damages? Learn more about the various ways AI may harm the environment, growing concerns among environmentalists and what its developers are doing to make a positive difference.

Excessive Energy Consumption in AI

One of the biggest concerns regarding AI environmental impact is how much energy these technologies use. While AI is still a relatively new tech, it’s already establishing itself as a major contributor to energy consumption and greenhouse gas emissions.

Advanced computers performing complex and powerful tasks require more energy than your laptop or smartphone. When OpenAI trained GPT-3, a model that processes and generates text, the project produced around 500 tons of carbon dioxide. That’s over 100 times more than the typical passenger vehicle emits a year—and GPT-3 is only one AI model.

One peer-reviewed study estimates that NVIDIA, a leader in AI computing, will ship 1.5 million AI servers annually by 2027. All of those servers running at full capacity use at least 85 terawatt-hours of electricity every year. That’s more energy than many small countries use annually. 

These figures add up quickly—and have real-world implications. Researchers predict the information and communications technology (ICT) industry will be responsible for 14% of global emissions by 2040. Such significant energy use must be a consideration for AI developers and users in the future.

Making AI More Energy Efficient

Fortunately, there are many people working to reduce AI’s energy consumption and improve its overall sustainability. Several viable methods can cut back on the technology’s carbon footprint, such as using existing models rather than training new ones or limiting how often you utilize them.

One solution with promise is energy capping. Researchers at MIT studied computers with limits on how much energy they were allowed to draw. While the cap slightly increased the amount of time tasks took to complete, it also reduced energy consumption by 12%-15%. Making AI algorithms more efficient by using energy caps or reusing existing models can trim the tech’s carbon footprint significantly.

AI’s Wider Climate Impact

Carbon emissions are only one facet of AI’s potential impact on the environment. While the technology itself is a heavy burden on energy systems, its practical applications may also lead to environmental issues. AI helps make various systems and industries more efficient—including those that have a negative impact on the planet.

In 2019, ExxonMobil partnered with Microsoft to use their cloud-computing software Azure, which relies on AI for performance analysis and other tasks. This partnership optimizes ExxonMobil’s fossil fuel operations, increasing oil production by 50,000 oil-equivalent barrels per day. More fossil fuels are a net loss for the environment’s health.

Many other industries that already take a heavy toll on the environment could increase their production thanks to AI. That means worsening their negative impact on the planet. Manufacturers that streamline their operations could increase deforestation, while more efficient marketing powered by AI-targeted advertisements will lead to increased consumption and waste.

Using AI to Streamline Sustainability

While some worry about the risk of AI being used to support environmentally damaging industries, others envision a future where AI powers sustainability. Rather than using algorithms to speed up fossil fuel production, environmentally conscious manufacturers can take advantage of AI to streamline greener practices.

Using relevant data, AI can create more efficient production and transportation processes that output fewer emissions. Companies may receive recommendations for how to reduce energy use while maintaining productivity, which helps save time, money and the planet. When it comes to agriculture, farmers can use AI to develop smarter irrigation systems and improve crop yields.

Organizations dedicated to environmental protection can utilize AI as well. This smart technology can predict natural disasters brought on by climate change, allowing greater preparation and faster response times. Everyone benefits when AI is used to combat climate change, whether through preventive measures or other mitigating applications.

The E-Waste from AI Technologies

As new technologies develop, the old ones are often thrown out. AI is no exception—and with such rapid growth, the turnover of AI hardware is faster than ever.

E-waste has been a growing issue since the first outdated computers were tossed into landfills. Such waste is predicted to reach nearly 75 million metric tons by 2030. AI requires a lot of hardware that could quickly join fax machines and floppy disks in the dump. Servers, chips, central processing units (CPUs), graphics processing units (GPUs), memory systems, circuit boards and other physical devices allow supercomputers to work their magic.

But when a faster, stronger, all-around better AI model comes along with shiny new hardware, the old stuff is thrown out. Complicating matters further is the fact that AI hardware, like a lot of computer parts, is not easily recyclable—especially when it contains rare earth metals.

Rare earth metals are crucial to how our modern tech functions, powering everything from smartphone screens to medical imaging machines and electric vehicle batteries. Unfortunately, recycling rare earth metals is a challenging task. The metals may blend together with other elements, making it difficult to separate the valuable parts for reuse—even with complex processes that use hazardous chemicals and a lot of energy.

Manufacturers have to get real about reducing AI-related e-waste if they want to lower their bottom line and improve sustainability. Fortunately, a solution may be found in the problem. AI algorithms can streamline recycling programs, potentially offsetting their own contribution to landfills. Automated, smart e-waste recycling operations are highly productive, yielding at least 90% accuracy—a promising advancement for AI-driven processes in the future. 

AI Environmental Impact: Risks and Opportunities

AI is already a powerful and sometimes controversial technology. While it presents potential risks and challenges for environmentally conscious leaders and users, there are seemingly endless opportunities for sustainable development. It’s up to us to use AI responsibly and make a positive impact on the planet with this brilliant tech.