AI in HR: A Helpful Tool, Not a Replacement for People

By Paulette McCormack  |  Fresh HR Insights

I’ve been thinking a lot about artificial intelligence lately, and not in the “the robots are coming for our jobs” way that fills the headlines. The question that actually interests me is quieter and, I think, more useful: where does AI genuinely fit into the work we do with people?

I recently sat down with a commentary by Dr John Howard, written for the American Journal of Industrial Medicine, on artificial intelligence and the future of work. It’s a few years old now, but the points it makes have aged well. What stayed with me wasn’t the prediction that machines will take over. It was how clearly it set out what AI can and can’t do, and why the human in the loop matters more than ever.

A tool, not a colleague

Howard makes a distinction worth holding on to. The AI we use today is “narrow”, very good at specific, well-defined tasks. The science-fiction version, a machine that genuinely thinks like a person, is still a long way off. So the honest question isn’t “will AI replace HR?” It’s “what should we let AI do, and what should always stay human?”

I use AI in my own work. It’s a brilliant first-draft generator, a fast way to pull a checklist together, a handy sounding board at eleven at night when a policy needs a once-over. For the routine and the low-risk, it genuinely helps me move faster. I’d be silly to pretend otherwise.

But there are two things it cannot do for you. And in HR, both of them matter enormously.

The first thing AI can’t do: be in the room

Real HR happens in conversations. A performance conversation that could go either way. A redundancy you have to deliver to someone whose mortgage depends on the answer. A business owner sitting across from me, frightened and out of their depth, who needs to know they’re not on their own.

AI can draft the letter. It cannot read the room, hold the silence, or carry the weight of what’s at stake for a real person. Howard’s commentary touches on this from the worker’s side too: the rise of “people analytics” and constant monitoring can leave employees feeling depersonalised and more stressed, not less. Technology pointed at people without judgement doesn’t build trust. People build trust.

The second thing AI can’t do: tell you when it’s wrong

This is the one I’d underline twice. AI will give you a confident, well-written answer whether or not that answer is correct. Howard describes the “black box” problem, where many AI systems can’t explain how they reached a conclusion, and warns about the “de-skilling” that creeps in when people lean on those systems instead of their own expertise.

Here’s the catch: to spot a wrong answer, you have to already know enough to recognise it. He gives a striking example from medicine, an AI trained to detect skin cancer only on light skin will simply repeat that blind spot rather than fix it. Garbage in, garbage out. The same is true in HR. A tool trained on generic or overseas content doesn’t understand the Fair Work Act, your modern award, or how the Fair Work Commission actually weighs a dismissal. It doesn’t know what it doesn’t know, and neither will you, unless you brought the knowledge to the table first.

That’s not an argument against AI. It’s an argument for keeping a real expert in the loop. The cost of “confidently wrong” in employment matters isn’t a typo, it’s an unfair dismissal claim, an underpayment, a redundancy done the wrong way.

Using AI to draft your contracts, policies or letters? Brilliant, just have them checked by someone who knows Australian employment law before you rely on them. A quick review now is far cheaper than fixing a problem later.

Where I’ve landed

AI is a superb assistant and a poor decision-maker. Let it take the first pass at the things that don’t carry risk. But the judgement, the “should we”, not just the “can we”, belongs to a human who understands your business, your people and the law that governs both.

The future of HR isn’t AI instead of people. It’s people, supported by good tools, doing the human parts better. I’m genuinely excited about that version. I just don’t want anyone mistaking a clever draft for the real thing.

If you’re weighing up where AI fits in your HR, or you want a human set of eyes on something it’s produced, that’s exactly what I’m here for. Have a question that hasn’t been answered?

QUICK FAQ

Can AI replace HR?"

No. AI is good at specific, well-defined tasks like drafting documents or pulling together checklists, but it cannot replace human judgement, read the dynamics of a difficult conversation, or confirm whether its own answer is correct. The best results come from people supported by good tools, not tools instead of people

Is it safe to use AI like ChatGPT for HR advice?

Use it with care. AI tools are often trained on generic or overseas information and may not reflect the Fair Work Act, your modern award, or the Fair Work Commission’s approach to issues such as dismissal or redundancy. Always have AI-generated HR content reviewed by someone who knows Australian employment law before you rely on it

Where can AI genuinely help with HR?

AI is useful for low-risk, routine tasks such as first drafts of documents, checklists, summarising information and acting as a sounding board. It helps you move faster, as long as a knowledgeable human checks anything that carries legal or people risk

What is the biggest risk of relying on AI for HR decisions?

That it produces a confident answer that is wrong, and you do not have the expertise to notice. In employment matters, a wrong answer can mean an unfair dismissal claim, an underpayment, or a poorly handled redundancy- costs that far outweigh the time saved

GoHighLevelAgency Arcade, About Us - Agency Arcade, Contact Us - Agency Arcade, Our Services - Agency Arcade