An AI system designed to flag possible tuberculosis on chest X-rays is now operating in about 40 healthcare facilities across six African countries. More than 90 clinicians have used the Nexus AI tool to evaluate over 25,000 scans, cutting reported turnaround times from weeks to under 45 seconds.
The deployment grew from Google Research work showing that deep-learning software could match experienced radiologists at detecting active pulmonary TB. Nexus AI converted those algorithms into a clinical product and obtained European Class IIb medical-device certification. It is deployed in South Africa, Sierra Leone and Côte d’Ivoire, with expansion under way in the Democratic Republic of the Congo, Zambia and Cameroon. Clinicians still make the final decisions.
One installation operates four kilometers underground at South Africa’s Mponeng Gold Mine, where silica exposure raises miners’ risk of lung disease. Portable X-ray equipment and offline AI let staff screen workers on site and refer suspicious cases for molecular testing during the same visit.
Google has also released adaptable medical models, including MedSigLIP and the open-weight MedGemma family. Indonesia’s health ministry is fine-tuning them on domestic scans so patient data can remain on local infrastructure. These systems support screening and triage; suspected cases still require confirmatory testing.