A recent study has shed light on the environmental impact of Microsoft-backed OpenAI’s language model, ChatGPT. The research reveals that for every 5 to 50 user prompts, ChatGPT consumes 16 ounces of fresh water due to the water-based liquid cooling used in massive data centers. This has led to a 34% increase in Microsoft’s global water usage in 2021-2022, primarily driven by AI research and the partnership with OpenAI.
The growing popularity of ChatGPT has raised concerns about the environmental implications of large language models. Massive data centers that power these models require significant amounts of water for cooling. Despite the attention given to energy-intensive practices like cryptocurrency mining, language models like ChatGPT have yet to face the same level of scrutiny.
The research conducted by Shaolei Ren and his team at the University of California, Riverside highlights the lack of awareness about the resource usage underlying ChatGPT. Ren stresses the need for users to be informed about the environmental impact of their interactions with the chatbot in order to make more sustainable choices.
In response to these concerns, Microsoft has pledged to monitor its emissions and work toward sustainability goals, including being carbon negative, water positive, and zero waste by 2030. OpenAI also acknowledges the energy and water-intensive nature of training large models and is actively working on improving efficiencies.
It is essential for the environmental impact of AI technologies like ChatGPT to be addressed and mitigated. As these models continue to gain popularity and usage, it is important for users, developers, and companies to consider sustainability and work towards minimizing their ecological footprint.
– AP News