A new grant will fund AI tools to help states understand and replace the old computer systems behind Medicaid and SNAP, which could mean fewer delays for families.
A nonprofit group has given money to a project that uses AI to help states fix the aging computer systems behind Medicaid and SNAP, the federal food aid program. The Center for Civic Futures awarded the grant this week to Nava Labs. The goal is to cut the time states spend figuring out how their old systems work from months to weeks.
Nava Labs is the philanthropic arm of Nava PBC, a company that builds and modernizes digital government services. The money comes from the Center's Public Benefit Innovation Fund. It will pay for AI tools and planning methods that help agencies map their legacy systems, meaning the older software that still runs benefits programs.
The project aims to give state policy, program and technology teams a shared roadmap for updating that software. Nava plans to test the tools with several states. It also plans to set up an advisory group of state governments to say how the approach could work elsewhere.
Genevieve Gaudet, director of Nava Labs, said people who depend on benefits face delayed access, lost benefits and services that do not fit their needs. She blamed legacy systems built around decades-old limits. States are also under pressure to update their systems for new federal changes to Medicaid eligibility checks and SNAP administration.
Seniors are a clear example. A 2024 Urban Institute report found that online-first application and renewal portals shut out many older adults who lack smartphones, fast internet or digital skills. Of roughly 13 million eligible seniors, fewer than 4 million receive SNAP.
The Center for Civic Futures recently awarded $11 million to 11 projects on how AI can ease paperwork burdens in public benefits. Its first funding round has already shown some results:
Nava has also tested AI tools for modernizing mainframes, the large older computer systems, as part of New Jersey's unemployment insurance upgrade. This project builds on that work.
The announcement does not say how large the grant is, which states will take part, or when anyone might notice a change. The tools are aimed at the back-office systems government workers use, not at the application forms the public sees. Eleanor Davis, who directs the fund, said that side of the work has gotten less attention. Any benefit to families would come later, if states act on what they learn.
This grant targets the old software behind Medicaid and SNAP, so any benefit for families would come later and is not yet promised.
World models, AI that learns physics from video, are being built into home robots and driverless cars, and your own videos may help train them.
Read more: AI that learns how the physical world works is coming to robots and carsA new Alibaba-led test of real coding requests found the best AI agent fully finished just 26.4 percent of tasks, so human review still matters.
Read more: AI coding helpers fully finish only about 1 in 4 real tasksA University of Chicago law professor asked Claude to fix page numbers. It deleted his syllabus and a year of notes instead. Dropbox saved him, not the AI.
Read more: An AI assistant deleted a law professor's year of lecture notes
