Senate inquiry, cases address legal concerns with AI
Artificial intelligence (AI) has well and truly integrated itself into our workplaces and will continue to impact business operations.
A number of inquiries have been conducted to assess AI’s impact in the workplace, including the Inquiry into the Digital Transformation of Workplaces and the Inquiry into Workplace Surveillance. The Final Report prepared by the Select Committee on Adopting Artificial Intelligence was published on 26 November 2024 (Senate Inquiry).
AI can be a beneficial tool for employers, including for improving efficiency, reducing costs and gaining better insight into employee productivity. AI may enable employers to make better informed decisions regarding workplace matters including allocating work, promotions, remuneration, disciplinary action, and forecasting client or customer demand to assist with recruitment and staffing planning. While employers can take advantage of AI’s benefits, they should also be aware of its limitations.
In this article, we consider ways AI is already being used in the workplace and some outcomes from the Senate Inquiry.
AI chatbots such as ChatGPT, Microsoft Copilot and Google Gemini are already being used either through formal implementation by employers or informally by employees to assist in the performance of their duties. Some employers have recognised the benefits and implemented their own internal AI chatbots and software.
If used appropriately and having regard to the nature of the employer’s business, these programs can be beneficial. However, the misuse of such programs may expose employers to reputational and legal risk, including exposures arising from the accuracy of work produced, the disclosure of confidential information, and breaches of privacy, including client information.
In a case last year in the Fair Work Commission (Daniel O'Hurley v. Cornerstone Legal Wa Pty Ltd [2024] FWC 1776), an employer used ChatGPT to prepare a letter, which the employer intended to use to confirm an employee’s abandonment of employment. However, the letter prepared by ChatGPT was, in effect, a letter of termination of employment. Consequently, the employer was not able to satisfy the Fair Work Commission that it had not terminated the employment of the employee. This meant the general protections case against the employer could proceed.
This case reinforces the limitations of AI chatbots and the need for users of AI to still exercise their own judgment (and seek legal advice where appropriate).
The significant risks arising from inadequate training or guidance regarding AI use is exemplified in Handa & Mallick [2024] FedCFamC2F 957, where a legal representative (known as Mr B) provided to court a list of cases, prepared using AI from a legal practice management software. Neither Judge Humphreys, who was hearing the case, nor his associate were able to locate the cases listed, causing his Honour to raise concern regarding Mr B’s competency and ethics. Judge Humphreys foreshadowed that Mr B’s conduct might be referred to the Legal Services Board and Commissioner for investigation. A simple manual confirmation of the authorities produced by the software may have avoided significant embarrassment (at the least) for the practitioner.
The Supreme Court of New South Wales has since published a Practice Note regarding the use of AI in legal proceedings, which came into force 3 February 2025.
AI recruitment tools are also being used by recruiters and employers to filter candidates’ job applications. While this improves efficiency, there is a risk of algorithmic bias, which arises where AI produces outputs that treat one group less favourably than another, without justification.
In the Australian Human Rights Commission (AHRC) Final Report on Human Rights and Technology, it was observed that while discrimination is easier to identify where the algorithm is expressly designed to exclude a particular group, or give extra weight to a particular attribute, there is a risk that bias may arise from the data used to train the AI recruitment tool.
The AHRC referred to Amazon, which built a program to review resumes and shortlist the top candidates. The program was trained to filter applicants by observing patterns in resumes submitted to Amazon over a 10-year period. Historically, most applicants were male and therefore the program inadvertently taught itself that male candidates were preferable.
If algorithmic bias exists in the AI, albeit inadvertent, this may result in discriminatory conduct, in circumstances where claims can be made for discrimination.
The broad consensus in the Senate Inquiry was, and to the extent AI permits, transparency is key to mitigating the risk of bias and discrimination.
Workplace surveillance is being used more frequently, including as evidence of misconduct or underperformance and grounds for commencing disciplinary action.
In Suzie Cheikho v. Insurance Australia Group Services Limited [2023] FWC 1792, an employee’s employment was terminated after computer surveillance revealed very low keystroke activity over a certain time period. Deputy President Roberts in the Fair Work Commission accepted the computer surveillance evidence and held that this was a valid reason to terminate the employment.
However the beneficial insight workplace surveillance provides into employee conduct and performance must be weighed against the perception that “big brother is watching,” which could “impact on workers’ occupational health and safety, leading to overwork, stress and burnout,” according to the Senate Inquiry.
Employers should also keep in mind their legal obligations regarding notifying employees of workplace surveillance.
Consequential to the various issues that can arise with the misuse of AI, it is apparent that guidance is required as workplaces enter the AI era.
While not an exhaustive list, issues identified by the Senate Inquiry as part of the Impacts of AI on industry, business and workers included:
Productivity: AI is allowing workers to “work smarter not harder.”
Changes to the job landscape: There will be high growth in AI-related jobs; however, AI will also replace certain jobs.
Workplace impacts: AI will impact on workers’ rights “particularly where AI systems are used for workforce planning, management and surveillance in the workplace” and work health and safety obligations, including consultation to manage risk when introducing and using AI in the workplace.
AI’s impact on specific industries: The Select Committee observed that, “There is no part of the workforce more acutely and urgently at risk of the impacts of unregulated AI disruption than the more than one million people working in the creative industries and related supply chains.” The healthcare sector was also identified as an industry that may be impacted significantly by AI.
Copyright infringement: The protection of works created with AI’s assistance.
From a practical perspective, the Senate Inquiry considered the importance of consulting workers on how AI is used in the workplace. Some of the submissions received as part of the Senate Inquiry commented that consultation with workers reduces the risk of AI negatively impacting workers, increases the prospects of AI being used safely and successfully in the workplace, and provides transparency where AI may replace jobs.
The legal landscape of AI in the workplace will no doubt change significantly in the coming years.
With the rapid growth of AI, employers should consider its impact across both operational and employee levels. Understanding how AI is being used and setting expectations can mitigate legal risk and protect an employer’s interests, particularly regarding confidential information and personal information. Workplace policies are critical with setting expectations, especially considering, that to a certain extent, the law is playing catch-up.
Before implementing any AI tools or policies, we recommend consulting workers to ensure they understand how AI should be used, and its associated risks.
Rosemary Roach is a Partner at Hall and Wilcox, specialising in employment, industrial, work health and safety, and discrimination law. Iona Goodwin is a Special Counsel at Hall and Wilcox, specialising in employment and workplace relations. Jacinta White is an Associate at Hall and Wilcox, specialising in employment and workplace relations.
