How Green is AI? Comparing the Carbon Footprint of Artificial Intelligence and Human Tasks

In recent years, Artificial Intelligence (AI) has made impressive strides, and its applications have spread to a variety of industries, including healthcare, banking, transportation, and environmental preservation. However, as the use of AI spreads, worries about its effects on the environment have surfaced, notably in relation to the energy needed to run and train AI models and the resulting greenhouse gas emissions. One of the most potent AI systems in use today, GPT-3, for instance, produces emissions during training that are comparable to those created by five cars over the course of their lifespan.

The environmental effects of numerous AI systems have been examined in a recent study, with a focus on their capacity to carry out tasks like writing and painting. A team of researchers has compared the emissions created by various AI systems, namely ChatGPT, BLOOM, DALL-E2, and Midjourney, with the emissions produced by humans when carrying out the same duties. Writing text and producing images are the two common tasks that have been highlighted.

The goal is to contrast the environmental impact of people performing these tasks with that of AI. The team has emphasized the interchangeability of humans and AI by demonstrating that these costs are typically lower than those paid when humans perform identical activities, notwithstanding the environmental costs associated with AI. The results have shown a startling discrepancy when it comes to creating words.

When creating a page of text, AI systems produce between 130 and 1500 times less carbon dioxide equivalent (CO2e) than a human would. This significant difference highlights the environmental advantages of AI in this situation. Similar to this, AI systems release 310 to 2900 times less CO2e than humans do when creating images. These numbers unequivocally show how much less emissions are produced when images are created using AI.

The team has shared that it is crucial to recognize that an emissions study by itself cannot provide a full picture as a number of important social repercussions and factors need to be taken into account, which are as follows –

1. Professional Displacement: In some industries, employment displacement may result from the use of AI to undertake jobs that humans have historically handled. It is important to properly handle this displacement’s potential economic and social effects.

2. Legality: It’s crucial to ensure AI systems are developed and utilized according to moral and legal principles. The legality of AI-generated content and its potential abuse must be addressed to avoid any harm.

3. Rebound Effects: When AI is introduced into different industries, it may have unanticipated implications that are referred to as rebound effects. These results could show up as higher use or production.

It is critical to understand that not all human functions can be replaced by AI. AI cannot do some tasks and positions that call for human creativity, empathy, and decision-making. However, the current research indicates that, compared to humans, AI has the potential to drastically reduce emissions in a variety of tasks. While these results are encouraging from an environmental point of view, they need also be taken into account in the context of more extensive ethical, economic, and societal factors to ensure that AI integration is consistent with shared objectives and values. The prospect of using AI to complete some tasks with significantly fewer emissions is a viable approach to solving current environmental problems.

Check out the Paper. All Credit For This Research Goes To the Researchers on This Project. Also, don’t forget to join our 30k+ ML SubReddit, 40k+ Facebook Community, Discord Channel, and Email Newsletter, where we share the latest AI research news, cool AI projects, and more.

If you like our work, you will love our newsletter..