Navigating AI in the Workplace: Risks and Opportunities for HR

This article is part of a series focussing on themes that will form part of the discussion at Kingston Reid's Future Workplace Law Summit 2026. The Summit will bring together experts from law, technology and business to explore the challenges and opportunities shaping tomorrow's workplace.

The Summit will be held over 2 days in Sydney on 3 – 4 September 2026.

Click here for tickets and further information.

In just a few short years, artificial intelligence (AI) has quickly moved from a somewhat novel tool to an everyday reality. Its increasing use in the workplace is prompting employers, regulators, courts and tribunals to reassess how existing legal frameworks apply to the use of these new technologies.

Here, we reflect on some of the key employment law issues arising from the use of AI at work, and the broader institutional response to AI’s growing role in workplace disputes.

There has been widespread adoption of AI across many aspects of the employment lifecycle: from recruitment, workforce planning, rostering, job allocation, workplace monitoring and productivity tracking, as well as more traditional HR activities like workplace investigations, performance and disciplinary processes.

While these tools can improve efficiency and support decision-making, over-reliance on AI to carry out these activities (or aspects of them) presents unique legal risks, especially where the “thinking” is outsourced to machines. Many workplace disputes turn on the reasons why a particular decision or action was taken. So, if an employer adopts an AI-generated recommendation without properly understanding how it was reached, they may find themselves responsible if relevant factors were overlooked, or the outcome is inaccurate, unfair or discriminatory.

For this reason, human oversight remains critical. Employers should ensure that AI-assisted decisions are reviewed and that the reasons for decisions are clearly documented. Of course, this can be challenging, as AI tools do not always make it clear how a particular response has been generated, and errors or gaps in reasoning are easily obscured.

The rollout of AI can be just as risky creating legal risks before it is even used. Most modern awards require employers to consult employees about major workplace changes that are likely to have significant effects on employees, which can include the introduction of new technology. This is not just limited to potential job losses; it can also apply where jobs are redesigned, duties change, or employees need to be retrained.

If an employer requires employees to use AI to carry out their usual tasks, or requires AI use to perform new tasks, this could trigger the consultation obligation. Failure to do so could invite a dispute before the Fair Work Commission, as affected employees seek to resolve their concerns with the umpire’s intervention. These dispute proceedings can slow down implementation and see employers forced to arrive at different outcomes than desired. Many trade unions (including the ACTU, CPSU, UWU and NTEU) have publicly expressed strong views about the need to consult workers about AI adoption, and so can be expected to intervene if they believe consultation obligations have not been met.

Surprisingly, AI usage can also be a work health and safety issue. In NSW, recent changes to work health and safety laws make it clear that employers’ duty of care extends to risks arising from “digital work systems”. While the state regulator is in the process of developing guidelines on the issue, there will no doubt be focus on associated psychosocial hazards, such as role overload or underload, reduced or poor job control, uncertainty about roles and responsibilities and inadequate consultation during changes, amongst others.

Of course, it is not just employers navigating this new space. Courts and tribunals are also adapting to the growing use and impact of AI. Earlier this year, the Fair Work Commission linked a dramatic rise in its caseload to a sharp rise in AI-assisted applications. An estimated 50,000 to 55,000 applications were expected to hit the Commission in 2025/26, with the final numbers yet to be released.

It is not surprising that the largest increase related to general protections and unfair dismissal applications, which had increased by 40% by December 2025. This can partly be attributed to employees’ access to free chatbots (such as ChatGPT and Claude), which makes navigating litigation processes and ‘legalese’ less challenging and, in some cases, assisted claimants in drafting seemingly credible submissions for what are actually unmeritorious claims.

To assist the Commission to manage this issue, the federal government has recently passed the Workplace Relations Legislation Amendment (Building Cooperative Workplaces No 1) Act 2026 (which you can read more about here).

These legislative changes have given the Commission greater ability to process claims more quickly, which includes dismissing those it deems to have little prospects of success at an earlier stage. Still, it is doubtful even these newly vested abilities will wholly end the deluge of “AI slop” being copied and pasted into the tribunal’s unfair dismissal application form.

The courts are also adapting to the increased use of AI in litigated proceedings. The Federal Court of Australia has published the Use of Generative Artificial Intelligence Practice Note, emphasising the need for users to understand AI’s capabilities, limitations and risks, and ensure its use does not adversely affect the administration of justice. The Court may also require parties to disclose their use of AI in proceedings.

Importantly, parties remain responsible for the accuracy of documents filed with the Court. While AI can assist with drafting, affidavits, witness statements and expert evidence must reflect the evidence of the person giving it. While the Court has not prohibited the use of AI technology in preparing these documents, it sternly warns against the presentation of false or inaccurate information. AI-generated content should therefore be carefully reviewed before being relied upon in any legal proceeding.

AI use must remain accountable: employers are still responsible for AI-assisted decisions and should maintain human oversight, clear reasoning and defensible records.

AI implementation can create separate compliance risks: consultation, change management and work health and safety obligations may be triggered by how AI is introduced and used.

AI is changing the workplace disputes landscape: as employees, employers and representatives increasingly use AI tools, courts and tribunals are responding with new expectations around accuracy, transparency and governance.

AI is attracting greater scrutiny: Courts and tribunals are responding to the growing use of AI in workplace disputes, reinforcing the need for accuracy, transparency and human accountability.

The views expressed in this article are general in nature only and do not constitute legal advice. Please contact us if you require specific advice tailored to the needs of your organisation.

Photo by Peter Rovder on Unsplash