Generative AI has become an incredibly powerful tool, redefining not just productivity but every aspect of our lives. However, while AI excels at day-to-day efficiency, its inefficiency in resource consumption cannot be ignored. According to The Sustainable Agency, entering a prompt in ChatGPT uses 10 times more electricity than a regular Google search, and every 20 to 50 questions entered costs approximately half a liter of water. From training new models to powering data centers, sustaining AI is fueling carbon emissions, producing excessive e-waste, straining natural resources and constantly releasing toxic compounds. As the technology becomes increasingly prominent, it is essential to confront its damaging environmental impacts.
Senior manager at Ubilinx Technology, Jonathan Chiang, frequently uses AI in his business to generate meeting summaries, format specification documents and search databases. Chiang acknowledges that rampant AI use is currently harming the environment. However, he considers his business relatively environmentally conscious. According to Chiang, his company is constantly innovating, employing AI to develop more energy-efficient semiconductors, for instance. With environmental, social and governance frameworks, which disclose businesses’ environmental data, further incentivising AI companies to take the environment into account, there is hope that AI innovation to support the environment will compensate for its detrimental effects.
Even as many AI companies, including Google and Microsoft, are gradually taking responsibility for their environmental impact, AP Environmental Science teacher Kyle Jones believes it’s very difficult for people to be consistently held accountable for their individual contributions. One of the most popular AI search engines, ChatGPT, developed by OpenAI in 2021, currently has over 700 million weekly users worldwide, and has been considered by many analysts to be the fastest-growing internet app ever created. With the unprecedented rise of these search engines, there has been a subsequent increase in electronic waste, with the equipment used in relation to AI being predicted to produce up to 5 million tons of trash by 2030.
For these reasons, Jones believes that the environmental damages caused by AI reflect a core issue in society, where individuals are so concerned with efficiency and comfort, they choose to ignore the long-term impacts.
“This is the same problem that we’ve run into over and over and over again with every single issue that exists when you have a lot of people who need to do something: no one cares enough to think about something every time they’re gonna do it,” Jones said. “Do you think about taking a shorter route every time you’re gonna get in the car and drive? Did you cut your showers down by one minute? The impacts of AI usage are just another thing that a lot of people don’t really feel the direct impact of, and so they’ll just keep doing their old habits because people are just trying to get through the day.”
Junior and Environmental Science Club officer Anirud Sainarayanan says people are not yet sufficiently aware of their individual AI use and its environmental impacts. Students are growing increasingly reliant on AI, even using it to write essays. Sainarayanan believes individuals need to evaluate the necessity of their AI usage to avoid wasting resources.
“People don’t realize that these data centers require a lot of water and energy, because for them, it’s just typing in prompts and getting back responses from these AI models,” Sainarayanan said. “For a lot of simple tasks, such as writing a very simple email, you don’t need AI because those are day-to-day things that you should be good at anyway. It’s also saving water and energy.”
Furthermore, Jones is a firm believer that people will struggle to take individual accountability for their actions, placing AI in the same category as other environmental problems, such as the overuse of electricity in homes. However, he finds that a system that emphasizes collective responsibility could encourage individuals to act. For example, with Senate Bill 100, California pledged to be 100% renewable and zero-carbon by 2045. Jones believes a similar regulation could protect the environment.
“You have to do a larger, top-down kind of thing,” Jones said. “We can’t expect every single person to take accountability. In an ideal world, sure, but that’s not the way the world works. It’s hard to get individuals to change their behavior in any long-lasting way. They might change it for a week, but then they’ll just go back. When you talk about the environment, you can’t really be optimistic for everyone to chip in and do their part for free.”
Chiang agrees with Jones, adding that environmental issues will likely face a resurgence in public attention as AI becomes increasingly prominent. He also stresses that if enough individuals are conscious of their environmental footprint — avoiding, for instance, saying “thank you” to chatbots — there can be a cascading effect, and more powerful organizations will take action.
“As the technology industry grows, we need a bigger community,” Chiang said. “We need more organizations, more nations and more countries to contribute to making it better. If you have that from top to bottom, then from the bottom spreading to the older people, from bottom to top, it could be a win-win strategy.”
Already, several initiatives have been launched to reduce the environmental impacts of AI use. Notably, FUHSD has already begun addressing AI’s harmful environmental impacts alongside other issues related to responsible AI use, with its August advisory lesson emphasizing that smart prompting correlates with more efficient messaging and lower energy consumption. Additionally, Sainarayanan recalls AI data centers being a subject of debate during an FUHSD Climate Collective meeting, with many attendees rejecting a hypothetical data center in Cupertino.
Chiang, Jones and Sainarayanan all believe that with more education and an emphasis on community action, people will realize the damage excessive AI usage can have on the environment, and efforts will be made to combat it while still reaping the benefits of AI research and innovation.
“We need everyone’s contribution,” Chiang said. “People will think, ‘OK, we’re just individuals. We cannot do much.’ But I would say that everyone has this responsibility. If we can invoke some impact, in the long run, it will be good for the individual and for society.”