Summary
Google.org has selected 15 academic institutions, social enterprises and nonprofits for a $30 million challenge focused on AI tools for public services. The projects will receive support from Google engineers and product experts and are intended to produce reusable open-source blueprints.
Google.org has selected 15 organisations for a $30 million Impact Challenge aimed at using artificial intelligence to improve public services. The recipients include universities, nonprofits and social enterprises working with health ministries, municipal agencies and other government bodies.
Google announced the selections on September 15, 2026, saying they were chosen from a global pool of more than 2,600 proposals. Each team will take part in a dedicated Google.org Accelerator, receiving pro bono support and hands-on guidance from Google engineers and product experts.
The projects are intended to produce human-centred AI tools and open-source blueprints that governments and other organisations can adapt, improve and share. The announcement describes projects spanning emergency response, healthcare, taxation, transportation, climate resilience and legal services.
AI projects targeting practical government problems
The selected projects focus less on general-purpose AI products than on specific operational problems faced by public agencies.
- Emergency response and infrastructure: CIV:LAB is working on faster coordination during heat, flood and infrastructure crises. The Disaster Mitigation Research Center at Nagoya University plans to map emergency medical supply and demand in real time during disasters. The University of Cincinnati Research Institute is developing infrastructure-failure predictions to help emergency teams protect services such as power grids and roads. The Illinois Association of County Engineers is focusing on infrastructure resource management and road safety in rural counties.
- Health and access to care: Flexbone is helping caregivers and patients find affordable outpatient care regardless of insurance status. Froncort.AI is guiding frontline health workers towards infectious-disease clusters involving illnesses such as tuberculosis, malaria and dengue. The Helium Health Foundation is working on early risk identification and access to specialised maternal healthcare. The University of Warwick is mapping disease risks and deploying mobile clinics for preventive screening.
- Transport and climate: Johns Hopkins University is helping municipal transportation departments make road-safety decisions. The Massachusetts Institute of Technology is combining real-time public-transit feeds with operational data to improve service delivery and rider communications. ICLEI Europe is developing tools for municipal staff to act as community energy help desks, while Khoch3 KlimaKarten is helping municipalities assess and implement climate-resilience projects more quickly.
- Government paperwork and legal services: Clear Letter is turning complex government letters into plain-language explanations and next steps. Code for America is working to speed free tax filing for low-income households. TOJIL is addressing case backlogs through prosecutor-supervised case drafting and transparent case tracking for victims.
Together, the projects illustrate several ways AI can be applied inside public systems: analysing changing data, identifying risks, helping staff navigate information and translating complex administrative material into more accessible language. The announcement does not specify a single AI model or technical architecture for the programme; the methods are expected to differ by project and public-service setting.
Why the open-source approach matters
Public agencies often face problems that are highly local: a transport department may need to combine its own vehicle feeds, while a county may need to manage a particular road network or emergency plan. Google.org says the projects will be developed as open-source blueprints, allowing them to be adapted and shared rather than treated solely as closed services for one organisation.
That approach could make work produced for one city, county or ministry useful to others with similar needs. It also places the emphasis on implementation support: recipients will work with Google engineers and product specialists while developing tools for public-sector use.
The announcement provides the selected organisations and their intended areas of work, but not project-by-project budgets, delivery schedules or evaluation metrics. The $30 million figure is for the overall challenge, and the announcement does not state how it will be divided among the 15 recipients. The programme’s practical results will depend on how each team integrates its tool with public data, staff workflows and local services.
The recipients are working across communities described by Google.org as ranging from Illinois to Lagos to Nagoya. The challenge therefore represents a portfolio of experiments in applying AI to government operations, rather than the launch of one unified public-service platform.