Trust at the core: How do we safeguard evidence-based healthcare in an age of AI?

AI will be one of the biggest disruptors in healthcare and medicine, particularly for systematic reviews. Systematic reviews are a type of research that uses a systematic, replicable methods to evaluate all available evidence on a specific question. They are built on established principles of rigor, transparency and reproducibility and are the are a cornerstone in evidence-based decision making for policy-makers, practitioners and others. The developers of many new AI solutions are stating they can do systematic reviews at the touch of a button, but this is marketing, and there is a risk that we will be flooded with low quality, non-systematic reviews. This is a concern as it could result in policies and practices, from local to international levels, being based on unreliable and biased research. This session will provide an overview of how the Cochrane Collaboration - an international, independent, not-for-profit organization that produces high-quality, trustworthy evidence (primarily via systematic reviews) - is investing in responsible AI implementation to ensure it is used in a safe and reliable way. The key takeaways are: 1. This is a challenge for the whole systematic review ecosystem, and we all need to commit and work together (as proposed by the Responsible AI in evidence SynthEsis (RAISE) initiative). 2. How to approach the decision to use AI in a systematic reviews, including considering the trade-off between introducing errors and the speed or cost-effectiveness. 3. What we need from AI tool developers to ensure AI use in systematic reviews does not undermine evidence-based decision making. 4. Global efforts are underway to align approaches for responsible AI for systematic reviews. Please note, the talk will focus on systematic reviews, but could apply across different types of research.