A new study from researchers at the University of Kansas found that when artificial intelligence (AI) writes and illustrates tasks, it emits less carbon than humans performing the same tasks.
While this sounds intriguing, researchers caution this does not mean human writers and illustrators should be replaced.
KU researchers found that AI systems emit between 130 and 1,500 times less carbon dioxide equivalent (CO2e) per page of text generated and about 310 and 2,900 times less CO2e per image than humans.
“We’ve had discussions about something that appears to be true in terms of AI emissions, but we wanted to look at the data and see if it truly is more efficient,” said Andrew Torrance, Paul E. Wilson professor of Law at KU. “When we did it, the results were kind of astonishing. Even by conservative estimates, AI is extremely less wasteful.”
Torrance compared established systems such as ChatGPT, Bloom AI, DALL-E2 and more when comparing it to human writers and illustrators.
How they did it
Researchers calculated the carbon footprint of a person writing using Energy Budget, a measure that considers the amount of energy used in certain tasks for a period. Like using word processing software per hour. This is multiplied by the average time it takes a person to write one page of text on average.
It is then compared to how much energy is used by CPUs that operate AI such as ChatGPT, which is considerably faster than humans.
For illustrations, results showed that DALL-E2 emitted about 2,500 times less CO2e than a human artist in the U.S. and 310 times less than an artist based in the less energy consuming nation of India.
Quality over carbon?
One of the main reasons that KU researchers believe carbon emissions should only be one factor to consider when comparing AI production to human output is due to the quality of the AI technologies as they exist now.
Often, ChatGPT and other programs are not capable of producing the quality of writing or art that humans can. As these technologies improve, that may change but as it stands now, there is no real comparison.
Another reason to consider is loss of employment causing substantial economic, societal and other forms of destabilization in the world using AI.
KU researchers said the best path forward is a likely collaboration between AI and human efforts, or perhaps a system where people use AI to be more efficient in their work while retaining control of the final products.
“We don’t say AI is inherently good or that it is empirically better, just that when we looked at it in these instances, it was less energy consumptive,” Torrance said.
That includes when it comes to legal issues where AI may be using copyrighted material in training sets. This may also be an instance where collaboration is a better fit with human and AI working together to weed out potential legal issues while also consuming less energy than if humans only wrote or illustrated a piece.
“This is not a curse, it’s a boon,” Torrance said of AI. “I think this will help make good writers great, mediocre writers good and democratize writing. It can make people more productive and can be an empowerment of human potential. I’m hugely optimistic that technology is getting better in most respects and lightening the effects we have on the Earth. We hope this is just the beginning and that people continue to dig into this issue further.”
