How Reliable Is Science?

During the last decade, there has been an emergence of isolated scientific studies that have been selectively cited to support claims apparently contradicting the broader evidence base, thus opening the door for scepticism on the one hand and “evidence-based” conspiracies on the other. This dangerous development in science and its wider perception within the population have been facilitated by the rapid growth and digital discoverability of the scientific literature, making it easier to search selectively for the publications that lend support to any, and often the most bizarre, hypotheses. This bias of selective reading of literature constitutes a major challenge for modern science and scientific communication. AI-based tools may help broaden evidence retrieval, yet their outputs require transparent validation and expert oversight to minimise selection bias and to ensure reliable interpretation.