AI is big business. But it’s also having a big effect on our carbon emissions: The storage, cooling, and processing power required to train a large language model can produce as much CO2 as driving over a million miles in a family car. This article explores sustainable AI, and how making strategic decisions early in the design process – as well as proper data management – can help save your enterprise a fortune, and potentially reduce CO2 emissions.

Key takeaways:

• Plan your cost and benefits to make sure you get the most out of your investment.
• Combine software engineering and data science principles for repeatability, efficiency, and simplicity.
• Start with an in-depth understanding of the data you’re using.
• Tune your AI models with tools and techniques for efficiency.