Generative AI and employment law: emerging trends and common pitfalls in the workplace

Generative AI and employment law: emerging trends and common pitfalls in the workplace

Introduction/background

The use of generative artificial intelligence (AI) has increased significantly internationally and within New Zealand in recent years, with wide-reaching implications in the legal and employment law sphere. Supreme Court Chief Justice Winkelmann reflected that the courts are operating during “a time of profound social and technological change”, one that sees the use of AI technology in New Zealand’s legal context increasing. This article discusses some of the AI trends being observed in the employment law space, common pitfalls of AI use in the workplace, and some practical tips that can reduce potential headaches for employers down the track.

What trends are being observed in the employment law space?

An increase in claims

A significant trend in the employment law space has been a steady uptick in claims, as barriers including often high legal costs and difficulties in formulating legal claims have been alleviated by AI. In 2025, the Employment Relations Authority saw a 12 percent increase in applications lodged compared with 2024, with personal grievances being the primary applications.[1] The Law Society has referred to the increased use of AI as a “tidal wave” of change that has the potential to help litigants access justice in a way they might not have been able to otherwise. In response to an influx of AI assisted claims, the courts issued practice guidelines in 2023 to provide a helpful framework for both lawyers and non-lawyers using AI in proceedings, applying to all courts and tribunals in New Zealand (including the Employment Relations Authority and Employment Court).

AI use by employees and a higher volume of materials submitted by claimants

AI tools are also increasingly being used by employees to assist them in navigating workplace disputes, including to:

• Determine if they have an employment related issue and researching possible remedies;
• Drafting written correspondence (including emails and letters), information requests and raising personal grievances;
• Providing often lengthy and detailed responses during disciplinary processes; and
• Drafting pleadings, submissions and witness statements for the Employment Relations Authority and Employment Court.

Notably, in addition to an increase in claims, the volume of material being submitted by claimants has also increased across the board. The Tenancy Tribunal has recently sought to draft additional guidelines on AI use following concerns about its use to generate incredibly lengthy submissions. In one example, a tenant submitted a 101-page report alongside 215 further pages of evidence, but after two hearings was only awarded $80 in compensation. The employment jurisdiction faces similar challenges, a trend that is likely to continue as the number of self-represented litigants and claims increases. Auckland Barrister Josh McBride also described seeing instances where litigants were producing disproportionate volumes of AI generated material for what the case entailed, with a potential benefit in “clearer rules being developed around the volume of AI-generated material people could file.” We are routinely seeing our employer clients receive often lengthy claims, with excessive use of legal jargon and where the relevant issues are not clearly defined.

Awareness of AI limitations – accuracy and hallucinations, inability to understand legal and cultural nuances

As AI usage has become more mainstream, there is greater awareness of its limitations generally and within the legal and employment contexts. The Court guidelines encourage users to be aware of these limitations, which include limited training on New Zealand law and court procedure and a lack of understanding of the broader New Zealand social and legal context (for example, an understanding of tikanga). AI chatbots also remain prone to “hallucinating” information and cases (generating information that is inaccurate or does not exist). In 2026 the Supreme Court issued warnings in two cases for the use of hallucinated authorities in submissions by self-represented litigants, stating that serious cases may amount to obstruction of justice or contempt of court, and even call for costs sanctions.[2] The Employment Relations Authority and Employment Court have also commented, reminding parties that it is their obligation to ensure information being relied on in proceedings is accurate.[3] AI hallucinations can even amount to misrepresentation – in a recent example, an employee who used Chat GPT to draft his CV without fully reviewing it was found to have misrepresented his previous employment.[4]

Pitfalls of AI use in the workplace to avoid

Employer decision-making

To meet their legal obligations in the employment relationship, employers must take care when using AI tools to assist in decision-making with respect to employees, including in restructuring decisions and disciplinary matters. AI should not be used as the decision-maker, including when determining disciplinary outcomes. In the case of disciplinary processes, failure to follow a fair process can result in grounds for unjustified disadvantage and unjustified dismissal. When making a decision that could result in the termination of employment, employers are obliged to genuinely consider explanations offered by employees and come to a fair and reasonable decision. Our view is that “outsourcing” this task entirely to AI could put employers in breach of their legal obligations.

Caution should also be taken when using AI in recruitment. Recent studies have found that algorithmic bias can lead to discrimination when used to screen candidates, and employers should maintain oversight to ensure that candidates are protected from unlawful discrimination, including by asking any questions that may indicate an intention to discriminate on one of the prohibited grounds covered by the Human Rights Act.

Maintaining privacy and confidentiality

The Privacy Commissioner has provided helpful information and guidance about the use of AI in the workplace, including potential privacy risks. Employers should be aware that any sensitive or confidential information entered into AI tools can be retained or disclosed by the provider of the tool, as well as for AI training purposes. The starting point is that businesses are expected to adhere to the Information Privacy Principles (IPPs) contained in the Privacy Act, which sets out the legal requirements for the collection, use and sharing of information. If employers misuse personal information as a result of reliance on AI then this unlikely to be considered a sufficient excuse under the law. Employers are expected to consider taking steps such as conducting a privacy impact assessment before using an AI tool, and ensuring human review prior to acting on AI outputs.

Practical steps for employers using AI tools in the workplace

• Avoid using AI tools to come up with a decision when dealing with employment matters and processes;
• Ensure a company policy on use of AI tools in the workplace is readily available, and includes topics such as protection of confidential information and quality assurance;
• Offer training of AI tools for staff, and have a process that enables users to provide feedback on AI use; and
• Be cognisant of the limitations of AI, including potential for hallucinations and limited understanding of New Zealand’s social, legal and cultural context.

As the use of AI increases worldwide and within New Zealand’s employment and legal landscape, it is helpful to be aware of the risks and practical steps that can be taken to safeguard your business. If you would like more information on navigating AI in the workplace, please contact a member of our national employment law team at Duncan Cotterill.

Special thanks to Partner Alastair Espie, Special Counsel Jeremy Ansell and Law Graduate Hela Hartdegen for preparing this article. 

Disclaimer: The content of this article is general in nature and not intended as a substitute for specific professional advice on any matter and should not be relied upon for that purpose.