Help needed: looking for real-world Machine Learning systems in government
A few weeks ago, I reached out to this network asking for compelling GenAI use cases in public-sector workflows. Not smarter chatbots, but examples of actual automation (or at least augmentation) of core government processes. The responses were thoughtful and generous, but many confirmed what I suspected: we’re still early.
That made me wonder about machine learning more broadly. I suspect there are more mature, functioning ML systems out there, ones that may not be flashy, but are actually working. But they may get limited visibility amid the GenAI hype.
So I’m turning to you again, this time with a different ask.
I’m currently working on a study about AI in public service delivery, and I’m hoping to learn from concrete examples of ML systems that are up and running in government. Not pilots or prototypes, but production systems with real-world outcomes.
If you know of projects that:
- Go beyond call center automation
- Inform predictive resource allocation
- Triage inspections, applications, or health services
- Detect fraud or assess eligibility
- Have been institutionalized — and ideally have metrics (time saved, fraud caught, benefits delivered, etc.)
…I’d be truly grateful to hear about them.
I’m genuinely interested in what’s working, what isn’t, and what made the difference. What helped something survive the pilot phase? What were the enabling conditions? What barriers had to be overcome?
Please drop a comment or message me directly. I’d be grateful for any leads, and I promise to share back what I learn.