AI is moving deeper into hospitals, from clinical decision support to ambient AI tools that help doctors document patient visits. Protege DataLab focuses on building real-world healthcare datasets and benchmarks to evaluate how AI performs in medical settings.
A viral 2026 post claims hospitals adopting AI faster recorded fewer in-hospital deaths. The post compares hospitals based on actual AI use in clinical notes rather than simply asking doctors whether they use AI.
The shared figures show mortality falling for both fast and slow AI adopters between 2023 and 2025. The post then claims a sharp gap emerged in 2026, with faster adopters reaching around 1,010 deaths per 100,000 patients.
However, these specific mortality figures have not been found in Protege DataLab’s publicly available research. Independent analysis also warns that the viral numbers should be treated as unverified, especially because the 2026 figures are described as projections.
That does not mean AI has no role in better healthcare. DataLab says its healthcare work uses real-world, multimodal patient data to create stronger AI evaluation benchmarks for areas such as oncology, cardiology, triage, and clinical support.
Protege also highlights a major challenge: healthcare AI needs reliable evaluation. Its researchers argue that flawed benchmarks can produce misleading conclusions about how well AI actually performs in clinical settings.
The bigger takeaway is clear: AI could transform patient care, but stronger evidence is needed before linking AI adoption directly to lower mortality. For now, the viral figures are promising, but not yet proven.
.png){kind=logo}
