will AI replace janitors?

Every single task in your job involves being physically present and responding to what's in front of you. A wet floor from a spill that just happened. A blocked drain. A light that's been left on in a room that should be empty. No software can do any of that. You're not processing data. You're moving through a building and fixing what's wrong with it.

The monitoring side of your work is underrated. You're the person who notices that a pipe is weeping before it becomes a flood, or that a fire exit is being propped open. According to O*NET task data, notifying managers about the need for repairs or additions to building systems is a core part of the role. That kind of on-the-ground observation, the kind where you know what normal looks like because you've cleaned this building every day for three years, isn't something you can automate.

And the chemical knowledge matters more than people think. Mixing cleaning solutions to the right specifications, knowing which cleaners damage which floor types, following procedures with power equipment so you don't strip a finish or void a warranty. That's trained judgment. It's not a checklist a machine ticks off. It's experience applied to a physical space where the consequences of getting it wrong are immediate and visible.

AI needs a digital interface to work. Your job doesn't have one. There's no text to generate, no image to classify, no document to summarise. A robot arm that could mop a floor would need to navigate furniture, avoid obstacles, handle variable surfaces, refill its own tank, and know when a floor is actually clean versus just wet. Boston Dynamics has spent years on robots that can walk across a room. Scrubbing grout in an irregular public bathroom is a different problem entirely.

Autonomous cleaning robots do exist. The Avidbots Neo scrubs large, open floor plans in warehouses and airports. The Tennant T7AMR does similar work. But both are limited to flat, obstacle-free surfaces with predictable layouts. They can't clean restrooms, empty bins, respond to spills, or handle anything that requires opening a door, bending down, or making a judgment call. They also cost between $30,000 and $80,000 each, which makes them a bad investment for most of the offices, schools, and hospitals where you work.

The liability issue is also real. If a robot misses a biohazard spill or fails to secure a building after hours, someone is responsible. That someone needs to be a person. Facilities managers aren't rushing to remove the human who can be held accountable and who knows the building.

Here's the honest picture: AI does almost nothing in your day-to-day work right now. The tools that are changing other jobs, things like ChatGPT for writing, or automated scheduling software, sit at the edge of your role at best. There are no AI tools that clean floors, empty bins, or inspect a building for hazards.

Where AI touches the janitorial profession at all, it's in facility management software used by the people above you. Tools like Hippo CMMS or UpKeep let facilities managers track maintenance requests, schedule cleaning rotations, and flag recurring issues. If your employer uses one of these, you might get work orders through an app instead of a paper list. That's the extent of it. You're not the one running the software.

Some large commercial buildings are trialling sensor-based systems where foot-traffic data tells cleaning staff which restrooms need attention most urgently. Companies like Infogrid sell building sensors that track things like soap dispenser levels or bin fill rates. In those buildings, you'd spend less time doing routine checks on areas that don't need them, and more time where the data says there's a problem. But the cleaning itself is still you.

For most janitors, the day hasn't changed much. You still arrive, work through your building section by section, handle whatever comes up, and leave it cleaner than you found it. The core rhythm hasn't shifted.

If you work in a large commercial building or a hospital, you might be using a digital task management app now instead of a paper sheet. Work orders come through your phone. You tick them off as you go. That's a small change in how you receive information, not a change in what you actually do.

What's spending more of your time isn't AI-related at all. Higher foot traffic in shared spaces post-pandemic means more frequent restroom checks and more reactive cleaning. The thing that hasn't changed, and won't change, is that the physical work is yours. No notification system or sensor grid changes the fact that someone has to be the one who actually shows up and cleans the thing.

The BLS projects 2% growth for janitors and cleaners between 2024 and 2034. That's roughly in line with average growth across all occupations, and it translates to about 351,300 job openings per year. Most of those openings come from turnover and retirement in a workforce of 2.4 million, not from new positions being created.

The growth is driven by demand, not by AI filling gaps. Buildings still need cleaning. Hospitals, schools, offices, and public spaces aren't going away, and the cleaning work inside them isn't automating. The sectors with the most janitor employment, healthcare and education, are both growing, which supports steady demand for this role.

The risk to job numbers isn't AI. It's budget pressure in local government and schools, where janitor positions sometimes get cut during funding shortfalls. That's a real and recurring issue. But it's a political and economic problem, not a technology one. The 72 out of 100 outlook score here reflects a job that's genuinely stable, not one that's safe only because people haven't gotten around to automating it yet.

The AI exposure score for this role is 0%, and it's likely to stay near zero for the next decade. For that to change, you'd need affordable, general-purpose humanoid robots that could navigate unpredictable indoor environments, handle liquids, open doors, and respond to unexpected situations. Companies like Figure and Agility Robotics are working on this, but they're years away from a robot that could do your full job, and the economics of deploying them at scale in schools and small offices don't work yet.

The more realistic near-term change is more sensor coverage and smarter scheduling in large facilities. Over the next five years, buildings will get better at knowing when they need cleaning, which means your time gets directed more efficiently. That's a change in how your work is organised, not a reduction in the work itself. A genuine threat to this role would require a leap in robotic dexterity and cost reduction that most researchers put at ten to fifteen years away, at minimum.

The clearest thing you can do is build toward the supervisory and facilities side of this work. Janitors who understand building systems, who can spot a maintenance issue before it becomes expensive, and who can manage a team are harder to replace and earn more. The Bureau of Labor Statistics puts the median pay for first-line supervisors of housekeeping and janitorial workers at around $47,000, compared to around $34,000 for janitors generally. That gap is worth pursuing.

Getting familiar with facilities management software is a practical step even if your employer doesn't use it yet. Knowing how to log work orders, track supply inventory, and communicate through a digital system makes you more useful as buildings adopt these tools. It also positions you for supervisor roles where you'd be the one assigning tasks through the platform.

Specialised cleaning certifications are another real differentiator. The International Sanitary Supply Association offers certifications like the Cleaning Management Professional credential. Healthcare facility cleaning is a separate specialisation with its own training requirements and higher pay, because the standards for infection control are strict and the consequences of getting it wrong are serious. If you're working in or near a hospital setting, that training path is worth looking into. The physical work is yours. The question is how much of it is skilled physical work, because that's where the pay and the job security actually sit.