2020
DOI: 10.1038/s42256-020-0219-9
|View full text |Cite
|
Sign up to set email alerts
|

The carbon impact of artificial intelligence

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
128
0
4

Year Published

2020
2020
2024
2024

Publication Types

Select...
9
1

Relationship

0
10

Authors

Journals

citations
Cited by 264 publications
(132 citation statements)
references
References 0 publications
0
128
0
4
Order By: Relevance
“…Matheson (2020) states that AI has become a sustainability issue focusing on specific ethical concerns, while Kobielus (2020) reveals that researchers have overestimated the benefits of AI and ignored its costs to raise the energy needed to make it operational. The data and computing power of AI comes at a cost that is approximately half a per cent of world energy consumption (Luccioni et al, 2020); AI, machine learning, information and communication technologies consume sizable electricity contributing to 8% of global energy use (Beyer, 2019); the carbon cost can be reduced by shifting to sustainable AI infrastructure (Dhar, 2020).…”
Section: Review Of the Literaturementioning
confidence: 99%
“…Matheson (2020) states that AI has become a sustainability issue focusing on specific ethical concerns, while Kobielus (2020) reveals that researchers have overestimated the benefits of AI and ignored its costs to raise the energy needed to make it operational. The data and computing power of AI comes at a cost that is approximately half a per cent of world energy consumption (Luccioni et al, 2020); AI, machine learning, information and communication technologies consume sizable electricity contributing to 8% of global energy use (Beyer, 2019); the carbon cost can be reduced by shifting to sustainable AI infrastructure (Dhar, 2020).…”
Section: Review Of the Literaturementioning
confidence: 99%
“…Moreover, AI-enabled personalized mobile marketing (e.g., Tong et al, 2020 ) and in-store communication and technology (e.g., Dekimpe et al, 2020 ; Grewal, Noble, et al, 2020 ; van Esch et al, 2021 ) can prompt unplanned offline purchases and impulsive buying, which, in turn, amplify consumption and its environmental drawbacks. Finally, information and communication technologies, applications, and systems related to AI themselves can have rebound effects caused by energy consumption and emissions of AI development, production, and deployment (e.g., Belkhir & Elmeligi, 2018 ; Dhar, 2020 ; Lange et al, 2020 ).…”
Section: The Ethics Of Ai In Marketingmentioning
confidence: 99%
“…After conducting a survey of 400 algorithms presented in research papers at two top AI conferences (IJCAI and NeurIPS), researchers reported that only 6% of the presented papers shared the algorithm's code, a third shared the data on which they tested their algorithms, and only half shared a partial summary of the algorithm (Gundersen and Kjensmo 2018;Hutson 2018). Several studies have investigated this issue in the context of energy consumption and carbon emissions (Lacoste et al 2019;Schwartz et al 2019;Strubell, Ganesh, and McCallum 2019;Henderson et al 2020;Dhar 2020). Indeed, after analysing a sample of 100 papers from the NeurIPS 2019 proceedings, Henderson et al (2020, 4) reported that none of them provided carbon metrics, only one of them "measured energy in some way, 45 measured runtimes in some way, 46 provided the hardware used" and 17 of them "provided some measure of computational complexity (e.g., compute-time, FPOs, parameters)".…”
Section: Normative Considerationsmentioning
confidence: 99%