Why this job splits down the middle
The work of a human resources assistant is built from two different materials. One half is paperwork: processing personnel action forms, keeping employee files current, posting openings, and setting up interview schedules. That work is structured, repeated and written down, which is where software and language models do their best work.
The other half is people. An employee asking why a benefit changed mid-year wants a straight answer from someone accountable. A new hire going through document checks needs a person who can look at the paperwork and sign off. A worker raising a problem with a manager is not going to open up to a chat window. Those conversations carry legal weight and personal weight at the same time.
Scale matters here too. The Bureau of Labor Statistics counted about 90,220 human resources assistants in the United States, with median pay of $50,610 (BLS, 2025). Its 2025 to 2035 projection has employment falling 6.2%. That is the pattern worth watching: not the job disappearing, but fewer seats as the clerical half of the role gets absorbed into HR systems. Entry-level hiring usually feels that first, which is the point of our guide to AI and entry-level jobs.
What software does, what it assists with, what stays with people
Start with the tasks AI can run end to end. Processing routine personnel action forms and updating records after a status change are now largely system steps with a human approval at the end. Drafting a job posting from a role description sits in the same group. The share of task time we put in that group is 40%. Our full view of how much of the job AI can handle today is explained on the coverage method page.
Then there is the assisted group, where a tool speeds the work but a person still owns the result. Screening applications against stated requirements is one: the model sorts and summarizes, the assistant checks the calls that matter and catches the ones the filter got wrong. Answering common questions about eligibility, leave or enrollment is another, with a drafted reply that an assistant corrects before it goes out. The share of task time in that group is 51%.
Last comes the work that needs a person in the room or on the line. In-person onboarding and employment eligibility document checks belong here, because someone has to look at the original and take responsibility for it. So does handling a complaint or a sensitive personal situation, where tone, discretion and knowing the people involved decide the outcome. The share of task time left to people is 9%.
What the evidence actually shows
Our evidence grade for quality parity on this job is D. That means there is no published, direct test of AI against trained human resources assistants doing their own tasks, so we give no parity number at all. Benchmarks on general office writing and document handling exist, but they do not measure HR record accuracy, policy interpretation or compliance work.
A test that would settle it is easy to describe. Give a model and a qualified assistant the same batch of real work: a set of applications to screen against a job description, a week of benefits and policy questions to answer, and a run of personnel records to update. Have HR professionals grade the output blind, and count errors on the items with legal consequences. Until something like that is published, treat confident claims in either direction with care. How the grades work is set out on the quality parity method page, and the wider approach is on our methodology page.
When the balance could shift
Most likely between 2035 and 2045 (8 in 10 of our scenarios). What that window measures, and how it is built, is explained on the replacement year method page.
Two things could pull it earlier. First, no hardware is needed: the robotics panel on this page shows the physical share of the work is minimal, so there is no robot to buy or install. Second, the capability arrives bundled. Employers do not run a procurement project to automate record updates; the feature appears inside the HR system they already pay for, and the clerical hours quietly shrink.
Two things hold it back. Record-keeping carries legal exposure, from eligibility documents to personnel files, and employers want a named person accountable for an error. And employees guard their own data. Pay, medical leave and disciplinary matters are handled differently when a colleague is involved, and companies know it.
What to do: ask your HR team which tasks the current system already automates, then volunteer for the ones it cannot touch.
How to stay needed in HR support
Lean into the duties that sit on the human side of the task list above. Own the onboarding experience, including the checks and conversations a new hire remembers. Be the person who handles sensitive cases, from a leave request tied to a health problem to an early complaint about a manager. And take the awkward policy questions, the ones where the written rule and the actual situation do not line up.
Two skills raise the floor under all of that. One is employment compliance knowledge: wage and hour rules, leave entitlements, record retention. The other is HR systems fluency, meaning you can configure workflows, audit what the automation produced, and explain its output to a manager. Our rundown of AI skills employers want covers the second in more detail.
If you are weighing a move, the nearest jobs are worth a look side by side. Human resources specialists handle recruiting, employee relations and benefits design with more judgment attached. Interviewers, except eligibility and loan share much of the intake and screening work. Payroll and timekeeping clerks are the closest neighbor on the records side. You can put any two of them next to each other on our compare page.
For context beyond one job, this role sits with other information and record clerks and in our administrative support sector view. If the BLS projection is what concerns you, our list of jobs AI is expected to shrink shows where this one falls among them.