GoKawiilMany researchers use AI tools for research, despite feeling that AI use is having a negative effect on science.Credit: Oscar Wong/Getty
When it comes to artificial-intelligence tools, scientists have a serious case of FOMO — fear of missing out. A Nature poll of more than 1,900 researchers uncovered concerns about the negative impacts of AI, but revealed that many people feel pressured to adopt the technology owing to concerns about being left behind.
Nearly half of poll respondents said that they feel broadly negative towards AI, with 63% saying that the risks of using tools such as large-language models (LLMs) to analyse data and scientific literature outweigh the benefits. Despite these concerns, nearly 60% of respondents felt that they would be left behind if they did not use AI tools in their work. The poll was posted on the Nature website, on social media and in the Nature Briefing e-mail newsletter.
“If you want to stay in the fight, you are left without a choice to adopt the use of some tools — particularly LLMs — to keep up with others who are using them,” says Alexander Gibson, a PhD candidate studying clinical prediction models at the Queensland University of Technology in Brisbane, Australia.
Gibson says that he has felt a “passive push” to use AI tools because they promise to make everything more efficient. However, he feels that, although AI models can produce work faster, the quality of that work is not always better — and is sometimes worse — than work done by humans. For example, Gibson says that he used an LLM to extract data from PDF articles for his research. This made his research feasible because extracting the data manually would have been challenging, but the LLM made mistakes. “The quality of the work had not improved.”
Opinions on AI
Nature asked researchers about their perceptions of AI, whether AI models are having a broadly positive or negative effect on science and how often they use these tools. The respondents came from 75 nations, with the highest percentage of answers coming from the United States (38%). Of the 1,907 scientists who responded, nearly 48% said that they feel negative towards AI. Meanwhile, 30% of respondents to the poll said that they feel positive about AI overall, and 22% were neutral.
Only 23% of the respondents felt that AI tools were having a positive impact on research, whereas 31% said that the technology was negatively affecting science. Nearly half of the researchers felt that the impact of AI use depends on how the tools were used. Models designed for specific scientific tasks tended to be more popular than were general-purpose LLMs such as those behind OpenAI’s ChatGPT chatbot.
Meet the academics refusing to use generative AI
Despite their scepticism, most poll respondents said that they use some form of AI tool in their research, with 25% stating that they use them every day. 26% use AI models weekly, 15% monthly and 34% said that they never use AI tools.
... continue reading