Jul 2 2020
In an industry first, Artificial Intelligence (AI) is being deployed to help review research papers and assist in the peer-review process. The state-of-the-art Artificial Intelligence Review Assistant (AIRA), developed by open-access publisher Frontiers, helps editors, reviewers and authors evaluate the quality of manuscripts. AIRA reads each paper and can currently make up to 20 recommendations in just seconds, including the assessment of language quality, the integrity of the figures, the detection of plagiarism, as well as identifying potential conflicts of interest.
"This is the first time AI this sophisticated has been made available to all participants of the peer-review process. AIRA can increase the quality of published science, maximize the efficiency of the publishing process and make peer-review more objective," says Kamila Markram, CEO of Frontiers.
Peer-review is widely recognized in the scholarly community as a vital process for ensuring quality in scientific literature. However, the demand for peer-review is growing at an immense rate. According to data from Dimensions, in 2019 more than 4.2 million articles were published, compared to just 2.2 million articles just 10 years prior. The increasing amount of scientific literature published and the growing demand for high-quality peer-review now makes the use of novel decision support technologies essential.
"Given the massive growth in the number of scientific manuscripts being published, we need to look to innovative new ways to ensure these manuscripts are evaluated rigorously, consistently, and efficiently. Artificial Intelligence allows for scale while at the same time upholding rigorous quality standards," said Daniel Petrariu, director of product development at Frontiers. "A good number of tasks, such as correcting language issues, identifying problems in figures, checking ethics statements, are time-consuming and can lead to reviewer fatigue. Others, such as checking for conflicts of interest between authors and reviewers or identifying plagiarism are only possible at scale with technological assistance. AIRA uses machine learning models to identify these types of issues to assist editors, reviewers, and authors and facilitate better editorial decisions."
Frontiers' Artificial Intelligence Review Assistant (AIRA) platform ensures manuscripts sent to peer-review reflect the standards that are essential for high-quality scientific research. AIRA is available to Frontiers' 100,000 strong editorial board, reviewers, and authors who can run automated quality checks on every manuscript submitted. By highlighting where critical decisions need to be made, this technology does not replace researchers but empowers them to make editorial decisions more effectively. It will support editors and reviewers by directing their attention to potential issues in manuscripts. These issues can then be addressed or clarified as part of the manuscript review. For authors, AIRA highlights issues in a standardized and objective way so they know exactly what needs to be addressed and why.
Daniel Petrariu added, "AIRA has been designed in collaboration with our Peer-Review team to evaluate multiple quality criteria on every manuscript submitted to Frontiers, as well as the various processes that take place during the peer-review process. It is a system designed to provide decision support through recommendations and semi-automated checks but ensures the final say is always made by a human with expertise in the relevant area. This user feedback is then captured and used for re-training and continuous improvement. It is this collaboration between machines and humans that allows us to achieve high accuracy but also efficiency."
Marie Souliere, head of publishing operations at Frontiers, said, "AIRA enables a level of sophistication beyond anything else in peer-review and truly is next-generation academic publishing at its most innovative. The major challenge the publishing industry faces is maintaining rigorous quality standards in the face of ever-increasing manuscript submissions - and AIRA enables us to do just that. By tagging potential issues that need to be addressed, AIRA empowers human experts to take editorial decisions more efficiently and effectively, reducing both the time to publication for authors and enabling the highest quality controls. We are very pleased to have been able to translate our peer-review and research integrity experience into AIRA technology to continuously improve and safeguard quality controls in scientific research publication."