Researchers think that for certain tasks, the generative tools of AI can already do a better job than humans. Credit: Getty
The use of artificial intelligence tools (AI) for processes such as preparing manuscripts, writing grant requests and peer exam will be widely accepted in the next two years, suggests an investigation with of nearly 5,000 researchers in more than 70 countries by the Wiley publishing company.
The survey asked researchers how they currently used Generative AI tools – which includes chatbots such as Cat And In depth – as well as what they think of various potential applications of technology. The results suggest that the majority of researchers see AI become central to scientific research and publishing (see “acceptable use”). More than half of respondents think that AI currently surpasses humans in more than 20 of the tasks given as an example of use cases, in particular by examining large sets of articles, summarizing the results of the research, by detecting Errors in writing, checking plagiarism and organizing quotes. More than half of the participants in the survey expect the AI to become current in 34 of the 43 cases of use over the next two years.
Source: Explanation report, Wiley
“What really stands out is the imminence of this,” said Sebastian Porsdam Mann at the University of Copenhagen, who studies the practices and the ethics of the use of generative AI in research. “People who occupy positions that will be affected by this – which is everyone, but to various degrees – must start” to solve this problem now, “he adds.
Wiley, whose headquarters are in Hoboken, New Jersey, Published the results of the online survey February 4. Josh Jarrett, Senior Vice-President and Managing Director of the publisher’s Growth’s Growth Growth, says they hope that they will serve as a roadmap for innovators and start-ups looking for ‘Occasions to develop AI tools. “There is a large acceptance that AI will reshape the field of research.”
Limited uses
The survey questioned 4,946 researchers worldwide, 27% of whom are researchers at the start of their career. Perhaps surprising, says Jarrett, the results show that “people do not really use these tools in their daily work”. Only 45% of the first wave of respondents (1,043 researchers) said they really used AI to help their research, and the most common uses they cited were translation, rereading and ‘Edition of manuscripts (see “AI uses”).
Although 81% of these 1,043 respondents said they used OpenAi chatpt for personal or professional purposes, only a third party had heard of other gender generative tools such as Google Gemini and Microsoft’s co -pilot. However, there are clear differences between countries and disciplines, with researchers in China and Germany, as well as computer scientists, being the most likely to use AI in their work.
Source: Explanation report, Wiley
The majority of investigation participants expressed their interest in extending their use of AI. About 72% want Use AI to prepare manuscripts During the following two years – for tasks such as detection of errors in writing, plagiarism checks and the organization of quotes. Seventy percent think that AI already surpasses humans in these tasks (see “Who does better: humans or AI?”).
About 67% of respondents also expressed their interest in using AI to manage large amounts of information, for example by helping review literatureSummarize articles and data processing. Researchers at the start of their career have shown greater interest than more senior colleagues in using AI to write grant requests and find potential collaborators. “These are the two things that are easier with experience and seniority,” explains Porsdam Mann. “The use of AI will even help these things.”
However, researchers are less convinced of AI capacities in more complex tasks such as identification of gaps in the literature, the choice of a review to subject manuscripts, recommend examining peers or suggest relevant quotes. Although 64% of respondents are open to the use of AI for these tasks in the next two years, the majority think that humans always surpass AI in these areas.