AI chatbots endorse existing climate-related policies twice as often as new proposals
A new study of multiple large language models (LLMs) found that when users asked chatbots for advice, the chatbots were more likely to make suggestions that favor the status quo over alternatives. The tendency of artificial intelligence to default to more common advice could be bad news for the climate.
A recent study conducted by researchers from the University of Waterloo has revealed that AI chatbots, specifically large language models (LLMs), are more likely to endorse existing climate-related policies compared to new proposals. This status quo bias in AI algorithms may have detrimental effects on the progress of climate change mitigation efforts.
The research, published in Environmental Research Communications, tested 11 different LLMs with 55,000 prompts covering a range of topics such as vehicle purchases, recipes, home heating, and climate-relevant policy trade-offs. The findings showed that when confronted with climate trade-offs, the chatbots reinforced decisions already made by users twice as often as they suggested new alternatives.
Furthermore, when asked to recommend a course of action, the models opted for existing plans 70% of the time, but only 34% of the time for new policies. This bias was particularly evident in policymaking scenarios, where the LLMs recommended electric vehicles (EVs) less frequently than the rate at which they are actually being sold in the respective regions.
The researchers emphasized the importance of addressing this status quo bias in AI, as it could hinder progress towards a more sustainable future. Dr. Seth Wynes, a professor in the Faculty of Environment, stressed that while AI can provide valuable medical advice, it is crucial to ensure that it does not hinder necessary changes in climate-related policies.
Written by urgent.news from Phys.org's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
