Google Research has outlined a series of artificial intelligence tools and partnerships aimed at expanding tuberculosis (TB) screening in areas with limited access to radiologists and imaging infrastructure.

The company says its earlier chest X-ray algorithms matched experienced radiologists on a 2019 benchmark, while subsequent evaluations included a 2021 assessment by the Stop TB Partnership and a prospective multi-site validation published in 2024. Google also says its partner, Pretoria-based Nexus AI, translated the algorithms into a clinical diagnostic tool that received CE MDR Class IIb certification and ISO 13485 quality-management certification.

According to Google, Nexus AI’s system is operating at about 40 healthcare facilities across South Africa, Sierra Leone and Côte d’Ivoire, with expansion underway in the Democratic Republic of the Congo, Zambia and Cameroon. The company says more than 90 clinicians have assessed over 25,000 chest radiographs, with screening results available in under 45 seconds. It also reports that more than 3,000 free screenings have been delivered at under-resourced public clinics.

Google has released chest X-ray embeddings, MedSigLIP and MedGemma through its Health AI Developer Foundations initiative. The company describes MedGemma as an open-weight multimodal model family that can process medical images and clinical text, including on local devices. Google says Indonesia’s Ministry of Health is fine-tuning the models on domestic radiographs for a national TB screening effort.

Separately, Google Research says its Health Acoustic Representations model, or HeAR, analyzes cough sounds recorded with smartphones and is being studied with partners in sub-Saharan Africa.