Human-caused increases in greenhouse gas levels in the atmosphere are a key driver of climate change. The information and communications technology (ICT) sector, as well as the data centre business, account for a sizable portion of worldwide greenhouse gas emissions.
As consumers and creators of digital technologies that run in data centers, we must do our share to reduce the carbon footprint of digital activities, thereby minimizing climate change.
To achieve that aim, it is critical that we recognize that even digital items take energy to manufacture and consume, and hence have a carbon footprint. This article helps you achieve that goal.
Furthermore, having access to factual information is critical as we discuss how to reduce our carbon footprint so that we can prioritize our efforts to address the issues that yield the greatest carbon savings. Training GPT-3 is predicted to have utilized 1,287 MWh and emitted 552 metric tons of carbon emissions.
We can estimate ChatGPT's daily carbon footprint to be 23.04 kgCO2e using the ML CO2 Impact calculator. The average Dane emits 11 tonnes of CO2e per year [7], therefore ChatGPT's daily carbon impact is around 0.2 percent of a Dane's yearly carbon footprint. ChatGPT's carbon impact over a year would be 365 * 23.04 kg = 8.4 tonnes, or nearly 76 percent of a Dane's yearly carbon footprint.