Work with Cochrane to help ensure that AI is monitored and used effectively
Monitoring and using AI effectively is a key focus for Cochrane and is a commitment that forms part of our new Scientific Strategy. We wanted to update you on our thinking and some of the areas we’ve been working on.
Recommendations for responsible AI:
We have now released the first draft of recommendations for responsible AI in evidence synthesis in collaboration with others invested in this cause. You can find it here, and it offers tailored advice for a diverse range of roles. Whether you're an evidence synthesist, methodologist, AI developer, or an organisation or publisher involved in evidence synthesis, this guidance is a first step to help clarify your responsibilities and alleviate some of your concerns around AI use. We’re now gathering feedback and will refine these recommendations into consensus-based guidelines during 2025. Please share your input through this survey and help us to shape the future of responsible AI in evidence synthesis.
Making sure that AI tools measure up:
The recommendations for responsible AI include a draft of comprehensive how-to guidance to help those using AI apply it responsibly. This includes best practices for developing and validating AI systems and emphasises the importance of transparency, rigour, and accountability. These elements are crucial in supporting you to make well-informed decisions about integrating AI into your work.
Cochrane hopes to formally open routes for AI tools and systems to be proposed and receive formal endorsement soon. To support this endorsement process, Cochrane is looking into setting up a group that will play a crucial role in guiding decisions about AI endorsement within Cochrane and potentially in the wider sector.
Developing and improving our tools using AI technology:
Cochrane is working to make our tools more effective by using AI and automation to improve currently lengthy processes. We’re exploring the use of AI in generating plain language summaries, identifying included studies with associated retracted papers, and improving search features to help users find the right information more easily in the Cochrane Library. We aim to adopt these technologies as quickly as possible without compromising on trust, making it simpler for users to find trusted information and for authors to get the guidance they need.
Investigating how AI is affecting evidence synthesis:
Those across the Cochrane community are actively studying the impact of AI on their work processes. Insights from the recent AI Symposium hosted by Cochrane Germany highlighted both the promise and challenges of AI in systematic reviews. The findings presented show that while traditional AI tools have seen only modest adoption in the sector, potential breakthroughs in language editing, data extraction and screening are being explored and look promising.
By working together, we can ensure AI becomes a powerful ally in the quest for high-quality health evidence. Cochrane Evidence Synthesis and Methods, our open-access journal, welcomes research from across this area to contribute to the evidence base on how we conduct evidence synthesis using AI and ensure we maintain our high-quality standards.