Jump to content

Environmental impacts of artificial intelligence

Add links
From Wikipedia, the free encyclopedia

The environmental impacts of artificial intelligence (AI) may vary significantly. Many deep learning methods have significant carbon footprints. Some scientists have suggested that artificial intelligence may provide solutions to environmental problems.

Carbon footprint of AI[edit]

AI has a significant carbon footprint due to growing energy usage, especially due to training and usage.[1][2] Researchers have argued that the carbon footprint of AI models during training should be considered when attempting to understand the impact of AI.[3] One study suggested that by 2027, energy costs for AI could increase to 85-134 Twh, nearly 0.5% of all current energy usage.[4] Training one deep learning model may use up to the same carbon footprint as the lifetime emissions of 5 cars.[1] Training and running large language models (LLM) and other generative AI generally requires much more energy compared to running a single prediction on the trained model.[5] Operating a model, though, may easily multiply the energy costs of predictions.[5] The computation required to train the most advanced AI models doubles every 3.4 months on average, leading to exponential power usage and resulting carbon footprint.[6]

BERT, a generative AI model trained in 2019, consumed "the energy of a round-trip transcontinental flight".[7] GPT-3 released 552 metric tons of carbon dioxide into the atmosphere during training, "the equivalent of 123 gasoline-powered passenger vehicles driven for one year".[7][8][9] Much of the energy cost is due to inefficient model architectures and processors.[7] One model named BLOOM, from Hugging Face, trained with more efficient chips and only released 25 metric tons of CO2.[8] Incorporating the energy cost of manufacturing the chips for the system doubled the carbon footprint, to "the equivalent of around 60 flights between London and New York."[8] Operating BLOOM daily was estimated to release the equivalent carbon footprint as driving 54 miles.[8]

Algorithms which have lower energy costs but run millions of times a day can also have significant carbon footprints.[8] The integration of AI into search engines could multiply energy costs significantly,[7][10] with some estimates suggesting energy costs rising to nearly 30 billion kWh per year, an energy footprint larger than many countries.[11] Another estimate found that integrating ChatGPT into every Google search query would use 10 tWh each year, the equivalent of the European Union's current residential energy usage.[10]

AI has caused both increased water and energy usage, leading to significantly more demands on the grid.[12] Due to increased energy demands from AI-related projects, a Kansas City coal-fired plant pushed back closing.[13] Other coal-fired plants in the Salt Lake City region have pushed back retirement of their coal-fired plants by up to a decade.[14] Environmental debates have raged in both Virginia and France about whether a "moratorium" should be called for additional data centers.[13]

In 2024, Google failed to reach key goals from their net zero plan as a result of their work with AI,[15][16] and had a 48% increase in greenhouse gas emission attributable to their growth in AI.[12] Carbon footprints of AI models depends on the energy source used, with data centers using renewable energy lowering their footprint.[6] Many tech companies claim to offset energy usage by buying energy from renewable sources, though some experts argue that utilities simply replace the claimed renewable energy with increased non-renewable sources for their other customers.[14] Analysis of the carbon footprint of AI models remains difficult to determine, as they are aggregated as part of datacenter carbon footprints, and some models may help reduce carbon footprints of other industries,[17] or due to differences in reporting from companies.[18]

Some applications of ML, such as for fossil fuel discovery and exploration, may worsen climate change.[3][9] Use of AI for personalized marketing online may also lead to increased consumption of goods, which could also increase global emissions.[9]

AI models with inefficiently implemented architectures, or trained on less efficient chips may use more energy.[7] Some skeptics argue that improvements of AI efficiency may only increase AI usage and therefore carbon footprint due to Jevons paradox.[17]

Water usage of AI[edit]

Cooling AI servers can demand large amounts of fresh water which is evaporated in cooling towers.[17][18] By 2027, AI may use up to 6.6 billion cubic meters of water.[19] One professor has estimated that an average session on ChatGPT, with 10-50 responses, can use up to a half-liter of fresh water.[17][20][21] Training GPT-3 may have used 700,000 liters of water, equivalent to the water footprint of manufacturing 320 Tesla EVs.[20]

One data center that Microsoft had considered building near Phoenix, due to increasing AI usage, was likely consume up to 56 million gallons of fresh water each year, equivalent to the water footprints of 670 families.[19] Microsoft may have increased water consumption by 34% due to AI, while Google increased its water usage by 20% due to AI.[21][6] Due to their Iowa data center cluster, Microsoft was responsible for 6% of the freshwater use in a local town.[21]

Other environmental impacts of AI[edit]

E-waste due to production of AI hardware may also contribute to emissions.[6] The rapid growth of AI may also lead to faster deprecation of devices, resulting in hazardous e-waste.[22] Some applications of AI, such as for robot recycling, may reduce e-waste.[23][24]

Climate solutions from artificial intelligence[edit]

Some climate scientists have suggested that AI could be used to improve efficiencies of systems, such as renewable-energy systems.[11] Google has claimed AI could help mitigate some effects of climate change such as predicting floods or making traffic more efficient.[16] Some algorithms may help predict the impacts of more severe hurricanes or help monitor emissions from sources.[9] One machine learning project, the Open Catalyst project, has been used to identify "suitable low-cost electrocatalysts" for battery storage of renewable energy sources.[3]

See also[edit]

References[edit]

  1. ^ a b Toews, Rob. "Deep Learning's Carbon Emissions Problem". Forbes. Retrieved 2024-07-04.
  2. ^ Heikkilä, Melissa (5 December 2023). "AI's carbon footprint is bigger than you think". MIT Technology Review. Retrieved 2024-07-04.
  3. ^ a b c "Achieving net zero emissions with machine learning: the challenge ahead". Nature Machine Intelligence. 4 (8): 661–662. 30 August 2022. doi:10.1038/s42256-022-00529-w. ISSN 2522-5839.
  4. ^ Erdenesanaa, Delger (2023-10-10). "A.I. Could Soon Need as Much Electricity as an Entire Country". The New York Times. ISSN 0362-4331. Retrieved 2024-07-03.
  5. ^ a b Desislavov, Radosvet; Martínez-Plumed, Fernando; Hernández-Orallo, José (2023-04-01). "Trends in AI inference energy consumption: Beyond the performance-vs-parameter laws of deep learning". Sustainable Computing: Informatics and Systems. 38: 100857. doi:10.1016/j.suscom.2023.100857. ISSN 2210-5379.
  6. ^ a b c d Sundberg, Niklas (2023-12-12). "Tackling AI's Climate Change Problem". MIT Sloan Management Review. Retrieved 2024-07-04.
  7. ^ a b c d e Saenko, Kate (25 May 2023). "A Computer Scientist Breaks Down Generative AI's Hefty Carbon Footprint". Scientific American. Retrieved 2024-07-03.
  8. ^ a b c d e Heikkiläarchive, Melissa (14 November 2022). "We're getting a better idea of AI's true carbon footprint". MIT Technology Review. Retrieved 2024-07-03.
  9. ^ a b c d Coleman, Jude. "AI's Climate Impact Goes beyond Its Emissions". Scientific American. Retrieved 2024-07-03.
  10. ^ a b Calvert, Brian (2024-03-28). "AI already uses as much energy as a small country. It's only the beginning". Vox. Retrieved 2024-07-03.
  11. ^ a b Kolbert, Elizabeth (2024-03-09). "The Obscene Energy Demands of A.I." The New Yorker. ISSN 0028-792X. Retrieved 2024-07-03.
  12. ^ a b Rahman-Jones, Imran (3 July 2024). "AI means Google's greenhouse gas emissions up 48% in 5 years". www.bbc.com. Retrieved 2024-07-04.
  13. ^ a b Piquard, Alexandre (2024-02-10). "'The explosion in AI-related electricity demand has already had local consequences'". Le Monde.fr. Retrieved 2024-07-03.
  14. ^ a b "AI is exhausting the power grid. Tech firms are seeking a miracle solution". Washington Post. 2024-06-21. Retrieved 2024-07-03.
  15. ^ St. John, Alexa (2 July 2024). "Google falling short of important climate target, cites electricity needs of AI". ABC News. Retrieved 2024-07-03.
  16. ^ a b "Google blames AI as its emissions grow instead of heading to net zero". Al Jazeera. Retrieved 2024-07-03.
  17. ^ a b c d Berreby, David (2024-02-20). "The Growing Environmental Footprint Of Generative AI". Undark Magazine. Retrieved 2024-07-04.
  18. ^ a b Berreby, David (6 February 2024). "As Use of A.I. Soars, So Does the Energy and Water It Requires". Yale E360. Retrieved 2024-07-04.
  19. ^ a b Hao, Karen (2024-03-01). "AI Is Taking Water From the Desert". The Atlantic. Retrieved 2024-07-04.
  20. ^ a b Danelski, David (28 April 2023). "AI programs consume large volumes of scarce water | UCR News | UC Riverside". news.ucr.edu. Retrieved 2024-07-04.
  21. ^ a b c O'Brien, Matt; Fingerhut, Hannah (9 September 2023). "A.I. tools fueled a 34% spike in Microsoft's water consumption, and one city with its data centers is concerned about the effect on residential supply". Fortune. Archived from the original on 8 February 2024. Retrieved 2024-07-04.
  22. ^ Zhuk, A. (2023-12-15). "Artificial Intelligence Impact on the Environment: Hidden Ecological Costs and Ethical-Legal Issues". Journal of Digital Technologies and Law. 1 (4): 932–954. doi:10.21202/jdtl.2023.40. ISSN 2949-2483.
  23. ^ Stone, Maddie (2022-02-03). "Can AI-powered robots solve the smartphone e-waste crisis?". The Verge. Retrieved 2024-07-04.
  24. ^ Shreyas Madhav, AV; Rajaraman, Raghav; Harini, S; Kiliroor, Cinu C (July 2022). "Application of artificial intelligence to enhance collection of E-waste: A potential solution for household WEEE collection and segregation in India". Waste Management & Research. 40 (7): 1047–1053. doi:10.1177/0734242X211052846. ISSN 0734-242X. PMC 9109239. PMID 34726090.
Retrieved from "https://en.luquay.com/w/index.php?title=Environmental_impacts_of_artificial_intelligence&oldid=1232856304"