Springer Nature and Google DeepMind Partner to AI-Assist Peer Review

Springer Nature is partnering with Google DeepMind to develop an artificial intelligence prototype designed to assist researchers in peer-reviewing academic manuscripts. The publisher announced on 26 October that the project aims to streamline the analysis of text and figures, intending to free up researchers to focus their expertise on questions requiring specialist scientific judgment, Research Professional News reported.

AI Integration into Peer Review Workflows

The prototype assistant is currently in testing and is designed to act as a support tool for reviewers rather than a replacement for human oversight. According to a statement from the publisher, the project seeks to create a more “rewarding reviewer experience” by helping researchers navigate complex manuscripts and identify specific sections that demand closer attention.

Saskia Steinacker, chief product officer for Springer Nature, emphasized that while the publisher is exploring how AI can support the growing volume and complexity of research, human judgment remains fundamental. “For us, innovation and trust have to go hand in hand,” Steinacker said. “That means testing what AI can add while being clear about where responsibility remains.” The initiative is part of a broader effort by the publisher to improve its internal workflows, which includes its existing manuscript submission platform, Snapp.

Challenges in Managing AI-Generated Content

While major publishers are moving to integrate AI tools, the broader academic community faces ongoing difficulties regarding the use of AI in research assessment. During a European Universities Alliances conference in Brno, Roisin Cheshire, head of individual-led research at Research Ireland, noted that funding bodies are encountering a rise in AI-generated reviews.

Cheshire stated that while organizations have implemented rules regarding AI, there is a struggle to police its use effectively. “We’re starting to see some reviews come in which are very obviously AI-generated,” Cheshire said. “How do you police that, and how do you say to that reviewer, we can’t use the review?” Cheshire added that although various agencies are launching pilot schemes, clear evidence that AI successfully simplifies the assessment process is “not there yet.”

Industry Approaches to AI Adoption

Springer Nature’s project follows earlier industry experiments with automated tools. In December 2023, the publisher Frontiers announced it was encouraging a “responsible, supervised” approach to AI in peer review. This decision followed an internal survey that indicated a small but increasing number of researchers had already begun using AI to assist with their review tasks.

Questions Regarding AI-Assisted Peer Review

What is the primary function of the Springer Nature and Google DeepMind prototype?
The prototype is designed to help researchers navigate academic manuscripts and identify areas that require closer examination, effectively streamlining the analysis of text and figures.

How do research funders view the impact of AI on assessment processes?
Research funders, including Research Ireland, have expressed concern that AI is not yet making the assessment process easier. Officials have reported challenges in identifying and managing AI-generated reviews as application volumes continue to rise.

Does the new prototype replace human peer reviewers?
No. The publisher stated that the project is intended to empower scientific experts while ensuring that human judgment remains at the center of the peer-review process.

More on Google: OpenAI Partners with Google Cloud to Bring GPT-6 Models · Google Updates AI Overviews: New Features for Official Links and Clickable Names