The Growing Role of AI in Workplace Disputes

In recent months, we have seen a distinct increase in the number of employees using artificial intelligence tools to help draft workplace correspondence. That includes drafting emails to their manager, letters raising concerns, responses during disciplinary processes, and, in some cases, Statements of Problem filed with the Employment Relations Authority. Communications that were once brief or informal are now often structured, lengthy, and framed in legal language.

There is certainly some value in this development. Not every employee is confident putting concerns in writing, and they may struggle to articulate themselves clearly. AI tools can help people organise their thoughts and present issues in a clearer way. This can result in issues being raised and addressed quickly and constructively.

That said, there are downsides. AI makes it easy to produce lengthy and technical correspondence with minimal effort. Employers can find themselves dealing with multiple detailed letters that escalate the tone of a situation before any direct conversation has taken place. Where communications immediately adopt a formal, adversarial stance, the employment relationship can come under strain very quickly.

Accuracy is another concern. AI systems can present legal information confidently but get it wrong. We have seen AI-generated correspondence that:

• misstates legal tests
• inaccurately summarises court decisions, and
• cites cases that do not exist.

These errors distract from the real issues and can make constructive engagement more difficult, particularly where the employee is emboldened to pursue claims based on misstatements of law by AI. We have seen this put significant strain on the employment relationship and ultimately result in worse outcomes for the employee.

This is a growing issue in the legal system more broadly.  In a recent decision, the Supreme Court of New Zealand criticised an applicant who had used AI to generate his submissions, as they included several hallucinated citations for cases that did not exist.  The Court noted the impact this has on the court’s processes and that applicants could face contempt proceedings where they misuse AI tools. According to an online database that tracks the misuse of AI in court proceedings, there have already been six cases in New Zealand since November 2024 where applicants have misused AI during proceedings.

Since 2022, there has been an increasing number of employees who are representing themselves in the Employment Relations Authority.  While the increase has been modest so far, we anticipate this may increase at a faster rate going forward given the widespread availability and adoption of AI tools.

For employers, AI can be useful to create initial drafts for documents and to organise information. However, it is still important for employers to understand their legal obligations and to seek professional advice where the situation escalates or becomes more complex.

None of this means AI should be viewed as a problem in itself. The key issue is how it is used.