By pillars
AI for livelihoods
AI that helps people find food, work, a home and the public services they are owed. What works, what is still a pilot, and what nobody has proven yet.
More stories
3 stories
Livelihoods · United States
Los Angeles County phones people at risk of losing their homes. A computer model decides who gets the call.
A machine-learning model built by the California Policy Lab at UCLA reads county records to find people at highest risk of homelessness, and case workers reach out before a crisis. Early results look promising, but they are not yet proven: a randomized trial is expected to report in 2027.
Food and farming · United States
A text helpline matches US families with the right free food pantry
Lemontree, a US nonprofit, uses a scoring model to weigh opening hours, entry rules, travel time and neighbors' reviews, then sends each family a short list by text or WhatsApp. It says more than 900,000 families have found free food this way.
Public services · Democratic Republic of the Congo
Can AI help spot hate speech on community radio? A pilot in eastern DR Congo
A 2025 pilot recorded and transcribed broadcasts from ten community stations in Goma and Bukavu, flagged possible hate speech for people to check, and answered with radio programs. No results are public yet, and a research phase with the Swiss university ETH Zurich runs to 2027.