One of the main promises of AI is its potential to help combat climate change. However, the current situation reveals a complex scenario when it comes to carbon emissions, particularly in the training and operation of large language models (LLMs). Sustainability experts are concerned about the increasing energy consumption as big companies use AI to justify their resource consumption.

The correlation between AI and climate change is more intricate than anticipated due to the high energy demands of AI models like LLMs. These models, such as those powering chatbots like ChatGPT, require significant amounts of power and computing resources compared to other search-based models.

The environmental costs and energy consumption associated with operating these models are considerable, with a substantial amount of electricity and water needed. The carbon footprint of these models can vary depending on the sources of electricity generation used by the grids powering the data centers where these models are stored and operated.

Key figures in the AI industry, including Sam Altman from OpenAI, Satya Nadella from Microsoft, and Jensen Huang from Nvidia, have expressed optimism about AI's potential to address climate issues. Despite these positive outlooks, the actual impact on the environment is more concerning, especially in terms of carbon emissions.

Google has committed to achieving net-zero carbon emissions by 2030, although this goal may involve the use of carbon offsets. The company's substantial carbon footprint and water consumption highlight the environmental impact of AI and related technologies. Microsoft has also pledged to invest in renewable energy sources to meet its green goals, acknowledging the significant energy consumption associated with AI operations.

While some experts acknowledge the rising computing and energy demands of LLMs, they argue that the benefits of using these models should be weighed against their environmental costs. It is essential to consider the potential positive impact of large AI models on the environment to justify their significant energy consumption during training and operation.