When ChatGPT was first released in late 2022, I was still completing my undergraduate degree in chemical engineering. For fun, my friends and I tried using it to solve some of our thermodynamics homework. The answer that ChatGPT spit out seemed correct at a glance — it used equations we had discussed in class and covered familiar topics. However, when we went through the solution line by line and compared it to the answer key, it was completely erroneous. I immediately became skeptical of ChatGPT’s usefulness for substantial tasks.
Nevertheless, as generative artificial intelligence (GenAI) has rapidly evolved over the past few years, it has changed how many of us operate. As users and programmers continue to train AI models and developers equip new generations with additional features, GenAI has become more accurate and useful for everyday tasks. Many organizations and companies, including AIChE, have offered trainings that teach their employees how to use AI to make the workplace more efficient. For many, AI has become indispensable. Unfortunately, as with all new technologies, GenAI (and AI as a whole) comes with a set of ethical dilemmas. In addition to privacy, transparency, and security fears, a new apprehension has risen among scientists and climate-aware individuals: the environmental impact of AI.
AI model training involves using repeated computations to adjust billions of parameters, which is a process that requires an immense amount of power. This process uses high-performance computing (HPC) infrastructure, which requires thousands of tensor processing units (TPUs), graphics processing units (GPUs), and central processing units (CPUs) that all run in parallel. This equipment is stored in data centers (Figure 1), which, on average, span roughly 100,000 ft2. However, these centers can be massive, with the largest U.S. data center (owned by Meta) spanning 4.6 million ft2(1).

▲Figure 1. Microsoft’s Chicago Data Center spans 700,000 ft2 and has a power capacity of approximately 198 MW. Image courtesy of CNET.
There are three primary...
Would you like to access the complete CEP Article?
No problem. You just have to complete the following steps.
You have completed 0 of 2 steps.
-
Log in
You must be logged in to view this content. Log in now.
-
AIChE Membership
You must be an AIChE member to view this article. Join now.
Copyright Permissions
Would you like to reuse content from CEP Magazine? It’s easy to request permission to reuse content. Simply click here to connect instantly to licensing services, where you can choose from a list of options regarding how you would like to reuse the desired content and complete the transaction.