As generative AI becomes more ubiquitous, concerns are growing about its environmental impact. Large language models (LLMs) are particularly energy-hungry, with some models producing up to 50 times more CO₂ emissions than others. The carbon footprint of LLMs is difficult to measure due to the lack of transparency from companies like OpenAI and Anthropic, but researchers are working to estimate the energy use of these models.
Forecast for 6 months: Expect increased scrutiny of AI companies’ environmental practices, with potential calls for greater transparency and regulation. Some companies may begin to disclose their energy usage and carbon emissions, setting a new industry standard.
Forecast for 1 year: As the carbon footprint of LLMs becomes a more pressing concern, we may see the development of more energy-efficient AI models and infrastructure. This could lead to a shift towards cloud computing and edge computing, reducing the need for massive data centers.
Forecast for 5 years: By 2028, data centers used for AI and other tech demands are projected to account for up to 12% of all energy in the US. In response, governments and companies may invest in renewable energy sources and energy-efficient technologies, reducing the carbon footprint of AI and other industries.
Forecast for 10 years: As AI continues to advance and become more integrated into daily life, we may see the emergence of new technologies that can mitigate the environmental impact of AI. This could include the development of carbon-neutral data centers, AI-powered energy management systems, and more efficient AI models that require less energy to train and run.
