Generational Artificial intelligence (AI) models like ChatGPT can boost patient flow and reduce paper use, but may not help hospitals avoid huge carbon emissions, according to research released at the COP29 climate conference in Baku, Azerbaijan, on Friday.
The study by Australian and UK researchers probed the impact of AI in healthcare using large language models. The paper, published in the Internal Medicine Journal, showed that ChatGPT can easily process thousands of patient records daily and substantially reduce resource consumption. But their energy consumption may be huge. Hospitals must think and use AI carefully as well as responsibly, including using shorter prompts to summarise patient data, the researchers suggested.
]"By the end of a hospital stay, it is possible to accumulate tens of thousands of words to your name. Unlike busy healthcare staff, private large language models similar to ChatGPT have time to read through and process this information," said Oliver Kleinig, who led the research from the University of Adelaide. "However, a single AI query uses enough electricity to charge a smartphone 11 times and consumes 20 millilitres of freshwater in Australian data centres. ChatGPT is estimated to use 15 times as much energy as Google," Kleinig added.
Kleinig noted that the environmental consequences of implementing large language models may be "very significant". More than the household emissions due to ChatGPT, "healthcare AI systems would likely have an even larger footprint" Further, the hardware needed for these AI systems requires extensive rare earth metal mining. This can potentially lead to habitat destruction. The manufacturing process alone can double the carbon footprint of AI operations, the team said. The study, however, suggests that AI could potentially reduce healthcare's environmental impact by improving patient flow and reducing paper use.
