New AI Tool Spots Burnout in Volunteer Coders Before They Quit

Much of the internet runs on free software written by volunteers. These "open source" maintainers have no boss, no HR department, and nobody whose job is to notice when they are drowning. Researchers built a tool called BurnRiSc that watches public GitHub activity — the site where coders share and track their work — and quietly calculates a monthly "Burnout Risk Score" for each person. It looks at 14 signals, including how often someone commits code, what hours they work, and how they write. Crucially, it compares each person only to their own past, not to anyone else.

The stakes are bigger than one person's bad year. When a lone maintainer walks away, projects can stall or break, and thousands of companies and apps depend on that code. In 2024, a security backdoor slipped into a widely used piece of software after its pressured maintainer handed control to a stranger. The researchers argue that catching burnout earlier could let companies and foundations step in with money, help, or extra hands before someone disappears.

The catch is that this is early, small work — 68 contributors across 10 projects — so it is nowhere near proven. It also raises uncomfortable questions. The tool reads public data, meaning it can screen people who never agreed to be screened, and a burnout score could easily become a reason to push someone out rather than support them. It also cannot tell burnout apart from a vacation, a new job, or one busy month.

So the real question is who ends up holding this tool. Used well, it could be an early warning system that protects the volunteers holding up critical infrastructure. Used badly, it becomes surveillance of unpaid workers who are already stretched thin. The paper doesn't settle that — but it shows the technology is now possible.