ChatGPT, a large language model for artificial intelligence, has been causing concern due to its massive carbon footprint. This is not surprising as AI systems require a considerable amount of energy and natural resources to function correctly.
The energy usage of ChatGPT is comparable to that of cryptocurrency mining, which has been known to consume an incredible amount of energy. An environmental researcher for an AI company revealed that training an earlier version of ChatGPT released as much carbon dioxide as a gas car driving over a million miles.
Creating large language models necessitates the manufacture of high-end chips that consume a lot of energy. Additionally, running these models on cloud servers necessitates a lot of energy use, resulting in a significant amount of emissions. Each time an individual inputs a prompt into ChatGPT, a computer consumes a lot of energy to complete the calculations required.
While Ethereum was able to reduce its energy consumption by more than 90%, Bitcoin still consumes a significant amount of energy to mine. According to Digiconomist, Bitcoin requires 95 terawatts of electrical energy each year. To put this in context, Kazakhstan uses an equivalent amount of electricity each year.
It is uncertain how much energy AI companies are using because they are not transparent about it. Experts predict that incorporating AI tools into existing products, such as search engines, will result in significant environmental consequences. However, AI applications can also be utilized for sustainable purposes, such as optimizing energy use, monitoring land and water usage, and assessing climate risks.
In conclusion, the energy consumption of large language models for AI such as ChatGPT is significant and requires attention. Improvements must be made to prevent adverse environmental consequences.
