Symbiosis in Science: AI and Humans Jointly Filter Medical Data

On March 22, 2026, the journal Scientific Reports published the results of the Screenathon 2.0 project, demonstrating the effectiveness of the hybrid approach (Human-in-the-Loop) in large-scale scientific research.

The IMPROVE consortium involved 27 specialists in the initial screening of 5,842 scientific articles related to patient-generated health data (PGHD). Using machine learning algorithms for preliminary clustering and text labeling allowed the team to complete a month's worth of work in just two days. This is important proof that AI in science acts not as an independent "black box" making final decisions, but as a powerful filter assistant that relieves scientists of routine cognitive load and accelerates meta-analysis.

Source: Scientific Reports / Nature