I use AI tools every day, and I’ve noticed a pattern in myself. Early on, I caught mistakes all the time. It would cite something that wasn’t accurate or make something up rather than admit it didn’t know. So I checked everything.
As the tools got better, I checked less. I’ve had deliverables where I looked at the output, thought “good enough,” and moved on without spot-checking everything. I work in marketing, so most of that risk is low. A miss in a piece of content is usually easy to fix.
HR doesn’t work that way. A confident wrong answer about pay or labor law can cost a business real money. It can also cost a manager the trust of the person sitting across the desk. And more of those conversations now start with someone saying they’ve already done their research.
When an Employee Uses AI to Research a Rest Break Claim
My colleague Nicole Blevins is an HR director at Xenium HR. She told me a story of an organization whose employee raised concerns about missing their 10-minute rest breaks, which Oregon requires.
The employee arrived with their mind made up. They said they knew their rights, had done their research, and were owed back wages and penalties for every break missed since they started. They added that they’d hate to file a BOLI complaint, since that would mean attorney fees for the company. Then they named an amount and offered to settle.
They also shared the research. It was clearly AI-generated and not made up. The information was valid. It applied to a different industry than the one in which this employee worked.
Nicole said that she listened and took in everything the employee presented. Then Xenium’s compliance team researched the question, spoke with the managers, and brought in a legal partner. The penalties the employee described didn’t apply to this employer. The concern about missed breaks still mattered and needed to be addressed.
How Missing Context Skews AI Answers on Employment Law
Nothing in the AI’s answer was invented. It responded to the question it was given, with the context it was given, and the context was thin. These tools don’t know your industry or your role unless you tell them.
Nicole raised another point that stuck with me. When someone frames a question in a leading way, for example, by stating that they missed their breaks and their employer owes them back wages and penalties, the tool tends to look for an answer that supports that view. In her experience, the free versions are especially eager to please. The more capable paid tools are more likely to stop and ask for the context they need.
I see the same gap in my own work. There’s a big difference between what I get on a paid plan and what’s available for free, and plenty of people are working from the free version. When I teach people how to prompt, I have them give the tool a role, provide clear instructions, include relevant context, and include an example or two. Then I have them ask questions if anything is unclear. If this employee had done that, they probably would have gotten a more complete answer, and the conversation with their employer would have started somewhere very different.
What to Say When an Employee Brings You AI Research
That still leaves the manager sitting across from someone who’s sure they’re right.
The natural reaction is defensive. Nicole described it plainly. The manager’s first instinct is that the claim is ridiculous; of course, they give breaks, and they certainly don’t owe all that money. The employee is feeling something just as strong from the other side.
Nicole has more than 10 years of experience in HR. Unless she’d missed something big, she knew the claim didn’t match the requirements. She still didn’t lead with that. Telling someone you’ve been doing this for a decade and they’re wrong is a fast way to end a conversation. So she asked questions instead. Tell me more about that. Where did you find this information?
Curiosity bought her team time to verify. It also kept the employee in the conversation. When they came back, the response followed a simple shape. Thank the employee for the research. Explain that you did your own, using verified sources and a legal partner. Name what’s valid about their information and where it stops applying. Then shift to what applies, which in this case was making sure the employee consistently received the breaks they deserved.
Really focus the conversation on how do we move forward rather than getting stuck in who’s right and who’s wrong.
Nicole Blevins, Director of HR Services, Xenium HR
Nicole was honest that this is a hard balance. Once HR comes back with findings that don’t support the employee’s claim, it can look like HR picked the employer’s side. An outside partner helps because the person doing the research isn’t the one who would pay the penalty. An internal HR person can sometimes play that role too, with its own complications. Either way, her answer is the same. Show that you heard both sides and took their research seriously before you responded.
When Managers Use AI to Answer HR Questions
Nicole told me she’s seen more of this in recent history, from employees and managers alike. Leaders at client companies come to her saying they looked something up with AI and plan to run with it.
That version concerns me more. An employee with a wrong answer usually surfaces in one conversation. A manager with a wrong answer can act on it before anyone thinks to check.
Nicole isn’t against using these tools, and neither am I. They’re powerful, and they make us more efficient. Her caution concerns nuanced situations where the right answer depends on details the tool lacks. As she put it, AI can be a good starting point, but it isn’t the final answer. When you’re not an employment lawyer, it’s hard to fact-check what a tool tells you about employment law. That’s why verification needs someone who knows the law, whether that’s an HR partner or legal counsel.
How to Use AI in the Workplace Without Overtrusting It
The point I keep coming back to came from Nicole. Having access to information and understanding it are two different things. AI gave this employee access. What it couldn’t supply was the judgment to know whether the information applied to them. Most managers don’t have that depth in employment law either, which is fine, as long as they know to check with people who do.
Just because you have answers and information doesn’t mean we’re applying, interpreting, or understanding it appropriately.
Nicole Blevins
For leaders, the practical version is short. Assume your employees are using AI and expect more specific, leading questions than you used to get. Don’t argue with the concern. Listen, gather everything they’re bringing you, and take it to the people who can verify it. Then come back with what applies and a plan for what happens next.
I still use these tools every day, and I’d encourage every leader to do the same. The advice I give everyone is advice I need too: fact-check everything before it goes out the door. When the answer touches someone’s pay or their rights at work, that check belongs with someone who knows employment law.
—
Brandon Laws is the host of Transform Your Workplace and the VP of Marketing and Product at Xenium HR, an outsourced HR and payroll partner for small and mid-sized businesses in Portland, Oregon.