Eradicating tuberculosis
TB is curable, yet it made 10.7 million people ill in 2024 and killed over a million; the old vaccine fails adults and millions go undiagnosed.
Open in the interactive tree →Tuberculosis is a bacterial lung disease that spreads through the air and can lie dormant in the body for years; nearly a quarter of the world's people carry the infection. Drug treatment takes months and resistant strains are rising. The only licensed vaccine, BCG from 1921, protects infants better than adults. Ending TB means finding every case, shortening treatment and protecting people before they fall ill.
As of October 2026
WHO's Global TB Report 2025 counts 10.7 million people falling ill and over 1.2 million deaths in 2024; only 8.3 million cases (78 percent) were notified, and available funding was US$5.9 billion against a US$22 billion annual target for 2027. The pipeline holds 29 drugs, 63 diagnostic tests and 18 vaccine candidates, six in phase 3; the lead vaccine M72/AS01E enrolled 20,000 people at 54 sites by April 2025 with no efficacy result yet, and Gates MRI and the Serum Institute of India agreed in July 2026 to start manufacturing ahead of the data. In early 2026 WHO announced diagnostic recommendations for near point-of-care molecular tests, tongue swabs and pooled sputum, and a real-life study in Pakistan reported 95 percent success for the 6-month BPaLM regimen against 74 and 78 percent for older regimens (preliminary, Epicentre).
What is missing
- A vaccine that protects adolescents and adults from lung TB (M72/AS01E is in phase 3, results pending)
- Cheap tests that find TB in children and in people who cannot produce sputum
- Shorter and safer treatment for both drug-susceptible and drug-resistant TB
- A reliable way to tell which dormant infections will turn into disease
- Stable funding: available money is about a quarter of the 2027 target
Becomes possible once solved
- An end to a disease that kills over a million people a year
- Fewer deaths among people with HIV, for whom TB is a leading killer
- Shorter courses and fewer people lost to follow-up
- Savings for health systems that now pay for long drug courses
Open steps
- A vaccine that protects adults Medium AI leverageFinish the M72/AS01E phase 3 trial and find better antigens and signs of protection, so adults and teenagers are protected against lung TB.
- Find every case High AI leverageScreen with AI chest X-ray and swab-based tests so the roughly 2.4 million people not notified each year are found, including children.
- Cure in four months or less Medium AI leverageTest new drug combinations that cure drug-susceptible and drug-resistant TB faster and with fewer side effects than today's six-month courses.
- Spot who will fall ill Medium AI leverageFind tests that show which of the many people with dormant infection will develop disease, so preventive treatment targets those at highest risk.
- Close the funding gap Low AI leverageRaise and keep funding for tests, drugs and vaccines: WHO counts US$5.9 billion available against a US$22 billion annual target for 2027.
Where AI could help
Medium AI leverage. AI already reads chest X-rays and predicts resistance from genomes, but vaccines and regimens need long trials and money.
- Reading chest X-rays to find TB in screening
- Predicting drug resistance from bacterial genomes
- Ranking vaccine antigens and drug combinations
Shown so far
- In June 2025 WHO announced that six AI software products for reading chest X-rays met its performance standards for TB screening in people aged 15 and older, after independent validation by FIND. source
- A 2019 study trained tree-based machine-learning models on 16,688 TB genomes and predicted resistance with AUC above 96 percent for amikacin, kanamycin, ciprofloxacin, moxifloxacin and multidrug resistance, using lab susceptibility tests as the reference. source
Prerequisites
- Vaccination1796
- Antibiotics1928
- Genome sequencing2003
- AI in medical diagnosis2018–2026