I’ve been thinking about a recent RNZ article discussing the growing impact of AI on employment disputes. (Yes, I am an RNZ-er!) And a conversation on Linked In about what motivates algorithms and LLMs – to be sure, it is more about dependency and clarity or success.
According to Rosemary Wooders, Partner at Bell Gully, AI is contributing to a significant increase in the volume and length of employment dispute documentation. Opinions, letters and strategies that once took considerable time to prepare and progress can now be generated in minutes. That sounds like progress, but it is creating significant additional pressure on the employment relations system in NZ. [linkedin.com]
“Unmeritorious” is the word the lawyers are using for all this guff. As a small business owner myself, particularly one in the HR space, this is a development worth paying attention to.
The New Workplace Reality
Not long ago, raising a workplace concern required a degree of effort. An employee had to gather their thoughts, write an email or letter, review it, potentially ask council or opinion from someone else, and decide whether to send it.
Likewise, employers who needed to respond to concerns, draft performance management documents, or navigate disciplinary processes had to invest time and thought into the process and paperwork.
Today, AI can do much of this drafting in seconds! Hey presto, an employee who feels frustrated after a meeting can ask AI to write a personal grievance letter and an irritated manager can ask AI to prepare a warning letter.
Both will likely receive something that looks polished, professional, and convincing, which is the challenge. Because while AI is excellent at generating words, it cannot determine whether those words accurately reflect the situation, or pass judgement on the wisdom, timing and context of submitting said letter.
When a Conversation Becomes a Document
One of the risks I see emerging is that AI makes it incredibly easy to escalate concerns. A misunderstanding that could have been resolved in a ten-minute conversation can suddenly become a three-page document. A workplace frustration can quickly be translated into legal language when an AI tool is prompted with the right terms. A difficult discussion can evolve into a formal dispute before either party has properly explored a resolution.
In many cases, the issue hasn’t become more serious, it’s simply become inflamed with a documentation process that included an LLM/AI tool. More documentation doesn’t necessarily mean more clarity, progress or resolution. Sometimes it just means more noise, risk and offense.
Confidence Isn’t the Same as Accuracy
One of the fascinating things about AI is how confidently it presents information. We have all seen it, right? I have seen Copilot tell me things that are entirely false with all the confidence and clarity of an expert. Of course, when I tell her she’s lost, she corrects course quickly and contritely, but if I didn’t know, I would have followed her lead.
Whether it’s writing a grievance, a response letter, or workplace advice, the language form your AI assistant sounds remarkably authoritative. But confidence should never be mistaken for correctness. Employment relations has always been highly dependent on context. The same behaviour from two employees may require completely different responses. The same complaint may have several valid perspectives. The same set of facts may support a range of potential outcomes.
AI doesn’t know your employees. It doesn’t understand your workplace culture. It doesn’t sit in meetings, observe interactions, or understand the history behind a working relationship. It is designed to simply predict what words are likely to come next. That’s useful, but it isn’t clever, motivated to help or wise to your context or story.
Is Good Faith Just Piffle?
I am no statistician, but I don’t expect AI is creating more workplace problems. I’m also convinced we are not finally seeing the problems that until now lay just under the surface. I do think we are in an interesting transition where AI is making workplace issues easier to formalise, escalate, and entrench. And I definitely think that is not a good thing.
One of the strengths of the employment relations framework in New Zealand has always been its emphasis on good faith. At its best, good faith encourages people to engage directly, share relevant information, and genuinely attempt to resolve issues before positions harden. AI doesn’t understand good faith. It understands prompts.
Give an LLM a grievance prompt and it will draft a grievance. Give it a disciplinary prompt and it will draft a disciplinary letter. Give it an argument and it will often help you strengthen it. What it won’t do is ask whether a conversation should happen before the letter is written. As small business owners, that’s probably the discipline we need to hold onto.
Plot a New Course
Before sending the AI-generated email, pick up the phone. Before getting your AI to generate a reply to a likely-AI-generated email, pick up the phone. Before responding with three pages, ask whether a ten-minute conversation could achieve more. Before assuming the worst, ask a few more questions. Use your authority as a business owner to de-escalate and choose to dial up connection and understanding.
The irony is that as artificial intelligence becomes faster and more capable, the value of human judgement increases rather than decreases. Businesses that thrive won’t be those producing the most impressive documents. They’ll be the ones that continue to build strong relationships, communicate clearly, address issues early, and apply wisdom when others are applying prompts.
At FixHR, we use AI every day. We’ll continue to use it. It is already making us more efficient and, in some areas, more consistent. But we also know its limits.
AI can help you write a letter. It can even help you build an argument. What it can’t do is discern whether that argument should be made, whether that letter should be sent, or whether a cup of coffee and an honest conversation would achieve a better outcome.
And if there is one thing I’d hate to see happen in our workplaces, it’s this: The easier it becomes to generate words, the less willing we become to have conversations. Because in employment relations, most disputes are not resolved by better writing; they’re resolved by better understanding.
If you are feeling yourself being sucked into the AI vortex, give us a call to help script a good strong, lawful, productive conversation for you to have with your staff member. You probably don’t need to be one of the statistics stuck in an overwhelmed dispute system, or spend your weekend fuming and prompting replies to an email you’ve received from your staff member this week.
Reach out to Ainsley HERE and focus on putting the human back in human resources.
