If you’re thinking about asking ChatGPT for legal advice about your general protections or unfair dismissal claim, be warned. The Fair Work Commission recently threw out the claim of a Sydney worker after discovering he’d relied on the chatbot to argue his case for unlawful termination.
On the advice of ChatGPT, the worker decided to lodge his general protections claim some two and a half years after his termination of employment. It was riddled with errors and failed to meet legal requirements. The Commission said the claim was “hopeless,” blaming the “obvious danger” of relying on AI for legal advice.
This story – Mr. Branden Deysel v Electra Lift Co. [2025] – is one of those Fair Work cases that remind us that while AI can be useful, it’s no substitute for professional advice.
Branden Deysel had ended his employment at Electra Lift Co on 19 October 2022. In 2025, he lodged a general protections claim with the Fair Work Commission under section 365 of the Fair Work Act 2009. He did this based on advice from ChatGPT. A claim under section 365 must be made within 21 days of dismissal.
Mr. Deysel filed his general protections claim almost two and a half years after his termination of employment, resulting in a delay of 919 days. Mr. Deysel argued that his was an unlawful termination. He alleged that Electra Lift Co. had contravened section 365 when it had dismissed him. The company, however, countered that the termination of employment was voluntary on Mr. Deysel’s part.
Mr. Deysel sought an extension of time to file his general protections claim, which the Fair Work Commission only grants in “exceptional” circumstances. He claimed the cause of the delay was “a lack of awareness” of his workplace rights. Mr. Deysel also claimed that Electra Lift Co. had contributed to him lodging the claim late. He said that he was concerned that if he lodged his general protections claim that the company would “take retribution against him.”
Mr. Deysel had admitted to the Commission that he had used ChatGPT to prepare his general protections claim. Electra Lift Co., meanwhile, argued to the Commission that the claim should be thrown out due to its late lodgement. The employer contended that this prejudiced the matter. It said that it should not be expected to defend against events about a termination of employment that took place over two years prior.
The Fair Work Commission held a conference in August 2025 to consider Mr. Deysel’s request for a time extension for his general protections claim. It took aim at Mr. Deysel’s use of ChatGPT to prepare his claim, which the Commission noted was obvious from its “deficiencies.” It was observed that his general protections claim did not “address the matters required” to prove his was an unlawful termination and that Electra Lift Co. had contravened section 365 of the Fair Work Act 2009.
The Commission observed that Mr. Deysel’s general protections claim included an extract from ChatGPT’s advice to him. ChatGPT had told him that Electra Lift Co. had violated several employment and other statutory obligations. It had recommended that he lodge a general protections claim. The Commission said that it could “see no basis for this advice.”
The Fair Work Commission also observed that ChatGPT had told Mr. Deysel to “consult a legal professional or union representative” to determine the appropriate course of action. However, the Commission noted that Mr. Deysel did not follow this advice. He had simply heeded the advice provided by ChatGPT and commenced his general protections claim.
The Commission highlighted the “obvious danger” of relying on AI for legal advice. It said that the end result for Mr. Deysel was that he lodged a claim that was “best described as hopeless.” This, the Commission noted, led to wasted time and money for itself and Electra Lift Co.
Turning to Mr. Deysel’s reasons for lodging his claim late, the Fair Work Commission found that they did not amount to the exceptional circumstances required to warrant a time extension. The Commission noted that it has in the past “repeatedly” said that ignorance of the statutory time limit is not a good reason for lodging a late claim. With regard to Mr. Deysel’s argument that he feared retribution from Electra Lift Co., the Commission said that this fear was not backed up by evidence.
Ultimately, the Commission found that the lengthy delay in submitting the general protections claim and Mr. Deysel’s use of Chat GPT resulted in him lodging an “unmeritorious” claim. It therefore refused to grant a time extension for his alleged unlawful termination.
This story, which caught the attention of media in June 2025, is a cautionary tale of using AI during the recruitment process. It involved 30-year-old Alexandra Frisby-Smith, who was dismissed from a potential job due to using ChatGPT. This happened while Ms. Frisby-Smith, a creative, systems and workflow consultant, went through the application process for a part-time position.
As part of the process, she was asked to complete a trial task within 30 minutes. The task involved developing a plan to streamline her potential employer’s daily operations. Ms. Frisby-Smith chose to use ChatGPT to refine her submission. She explained that she started “brainstorming” the best ways to streamline operations. Once she worked out an overview of how to do it, she then “popped that into ChatGPT.”
Ms. Frisby-Smith said that ChatGPT created something that was a “beautiful” representation of what she was thinking. She argued that this was the most effective way to complete the task, given the short time limit. But after sending it to her prospective employer, Ms. Frisby-Smith received a rejection email. It said that she’d been dismissed from the recruitment process due to “heavy use of ChatGPT.”
On social media and while speaking to the press, she defended her use of AI. Ms. Frisby-Smith said that she was “proud” to have used ChatGPT, arguing that it’s the “most efficient way of working ever.” She criticised the widespread negative perception of AI and asked when employers and society at large will “stop shaming people who use it?”
Ms. Frisby-Smith pointed out the hypocrisy of the employer who rejected her. She pointed out that they had noticed the formatting of her task was typical of AI. She therefore concluded that they “obviously use it as well…so why are you shaming me.”
While use of AI is landing some workers in hot water, it’s doing the same for some employers. No more so than consulting powerhouse Deloitte, which in August 2025 faced intense public pressure and political scrutiny after a $440,000 report it produced for the federal government was found to be riddled with AI hallucinations. The firm has been forced to issue a partial refund to the Albanese government.
Published in July 2025, the 237-page report was created for the Department of Employment and Workplace Relations. It intended to provide an expert independent review of the IT system the Department uses to automate penalties for jobseekers.
The media storm around the report erupted in August 2025. This was when University of Sydney academic Dr. Christopher Rudge publicly revealed that it featured around 20 fabricated references and citations. These hallucinations included citations for non-existent books and research papers, including one wrongly attributed to a University of Sydney law professor. The report also had an incorrect reference and a fabricated quotation attributed to a Federal Court judge.
After the errors were made public, Deloitte acknowledged that some footnotes and references were incorrect. However, it maintained that the AI hallucinations didn’t affect the bulk of the report’s findings and recommendations. The firm republished the report with an appendix that mentioned it had been created with the help of Azure OpenAI GPT – 4o.
AI is not only getting workers and employers in hot water but is also threatening the jobs of thousands. In 2025, a raft of AI-driven redundancies took place at some of Australia’s biggest employers. This included Telstra, NAB, ANZ, Westpac, Bendigo Bank and the Bank of Queensland. These companies have not generally cited AI as the direct cause for, in some cases, the termination of employment for hundreds of staff. However, the efficiency achieved from the technology is widely considered a major contributing factor.
The most explicit admission came from the Commonwealth Bank, which in July 2025 revealed it had axed 45 customer service roles. The company said this was due to the introduction of an AI-powered “voice bot” that replaced workers. A former Commonwealth Bank employee told media that roles at the company are being “hollowed out and recast” as AI is embedded into workflows.
Another ex-employee, Kathryn Sullivan, who had been with the Commonwealth Bank for over 25 years, told media that she had worked alongside the bank’s AI chatbot. In the years leading up to her termination of employment, she worked in the bank’s customer messaging team. She was responsible for developing scripts and testing responses for its AI chatbot, known as “Bumblebee.”
After helping refine the chatbot, Ms. Sullivan said she was in “complete shock” when she was told her role was redundant. She described the brutal realisation that she had “inadvertently” trained a chatbot that “took my job.” She had expected to be redeployed once her work on the chatbot was complete, not dismissed.
The emotional impact of her experience was highlighted when Ms. Sullivan confronted the CBA’s chair and CEO at the bank’s annual general meeting. She asked them what specific measures were in place to safeguard current staff from being displaced by AI and offshoring.
The twist in this story came around a month after the redundancies, in August 2025. The Commonwealth Bank was forced to backtrack on the redundancies and apologise, calling its initial decision an “error.” This embarrassing reversal came after the Finance Sector Union raised a dispute. The union and affected workers successfully argued that the chatbot hadn’t allowed the bank to reduce its workload by 2,000 calls a week, as it had claimed.
In fact, the bot actually led to rising call volumes. This forced management to offer overtime and pull team leaders onto the phones to cope with the surge. At the Commonwealth Bank’s annual general meeting, its chair and CEO publicly admitted to the “mistake” and an “unnecessarily challenging process” for the affected staff. The bank conceded that the 45 roles it had cut were “not redundant.” It also admitted that it did not factor in “all relevant business considerations” when declaring the roles redundant.
The backflip was hailed as a “massive win” for workers, highlighting the importance of union advocacy in holding corporations accountable. It demonstrates that while the threat of AI to jobs is real, its current capabilities are often overhyped by corporate leaders looking to cut costs.
If you’ve been treated unfairly at work, Australian Workplace Discrimination Representatives is here to help. With over 30 years of experience, we’ve guided thousands of Australian employees through general protections, discrimination, bullying, redundancy and sexual harassment claims.
Our team provides expert guidance to help you understand your rights, prepare your claim, and achieve the best possible outcome. Strict Fair Work time limits apply to make a claim, so don’t wait. Contact us on 1800 333 666 for a free and confidential consultation today.
