"Green consumerism," is the idea that consumers' decision to purchase or not to purchase an item is mediated at least partly by considering the item's impact on environmental or social conditions.9 In response to a general upswell of awareness around sustainable development goals (SDGs) and climate change, large corporations such as Google and Amazon have provided users with tangible sustainable choices. For example, when checking out on Amazon, if a customer has more than one item in an order, they can opt into Amazon Day, which delays shipping until all items are available to be placed in the same shipping box. Additionally, Amazon allows users to filter a search to only show items containing a climate pledge certification. Similarly, Google's search engine allows users to filter for low-emission options when searching for flights, as well as enabling users to search eco-certified hotels when booking travel plans.

Barroso's seminal contributions in energy-proportional warehouse-scale computing launched an era where modern datacenters have become more energy efficient and cost effective than ever before. At the same time, modern AI applications have driven ever-increasing demands in computing, highlighting the importance of optimizing efficiency across the entire deep learning model development cycle. This paper characterizes the carbon impact of AI, including both operational carbon emissions from training and inference as well as embodied carbon emissions from datacenter construction and hardware manufacturing. We highlight key efficiency optimization opportunities for cutting-edge AI technologies, from deep learning recommendation models to multi-modal generative AI tasks. To scale AI sustainably, we must also go beyond efficiency and optimize across the life cycle of computing infrastructures, from hardware manufacturing to datacenter operations and end-of-life processing for the hardware.

Artificial intelligence has become the tech industry's shiny new toy, with expectations it'll revolutionize trillion-dollar industries from retail to medicine. But the creation of every new chatbot and image generator requires a lot of electricity, which means the technology may be responsible for a massive and growing amount of planet-warming carbon emissions. Microsoft Corp., Alphabet Inc.'s Google and ChatGPT maker OpenAI use cloud computing that relies on thousands of chips inside servers in massive data centers across the globe to train AI algorithms called models, analyzing data to help them "learn" to perform tasks. The success of ChatGPT has other companies racing to release their own rival AI systems and chatbots or building products that use large AI models to deliver features to anyone from Instacart shoppers to Snap users to CFOs. AI uses more energy than other forms of computing, and training a single model can gobble up more electricity than 100 US homes use in an entire year.

Marc Andreessen famously said that software is'eating' the world, and now we have AI eating up software. However, in this original formulation, the'world' represented the economic slice of the world: How businesses operated and the profits they made were the core concern. With the push towards the triple bottom line, where all three Ps--profit, people and the planet--are taken into consideration, we must re-examine how AI is eating our planet! AI systems are very compute-intensive, i.e., their design, development, and deployment consumes a lot of cycles on a computer, typically utilising one or more Graphical Processing Units (GPUs). With the prevalence of cloud computing, we now have most training and inference jobs for these systems running in large data centres, that in turn have a rising carbon footprint.

This piece was the winner of the inaugural Gradient Prize. AI systems are compute-intensive: the AI lifecycle often requires long-running training jobs, hyperparameter searches, inference jobs, and other costly computations. They also require massive amounts of data that might be moved over the wire, and require specialized hardware to operate effectively, especially large-scale AI systems. All of these activities require electricity -- which has a carbon cost. There are also carbon emissions in ancillary needs like hardware and datacenter cooling [1]. Thus, AI systems have a massive carbon footprint[2].

AI systems are compute-intensive: the AI lifecycle often requires long-running training jobs, hyperparameter searches, inference jobs, and other costly computations. They also require massive amounts of data that might be moved over the wire, and require specialized hardware to operate effectively, especially large-scale AI systems. All of these activities require electricity -- which has a carbon cost. There are also carbon emissions in ancillary needs like hardware and datacenter cooling [1]. Thus, AI systems have a massive carbon footprint[2]. This carbon footprint also has consequences in terms of social justice as we will explore in this article.