Skip to main content
NOPE
Medium reported severity Lawsuit Dismissed

Fair Work Commission in Sydney upholds a software developer's dismissal, finding his reliance on AI-drafted complaint emails was counterproductive and led to it

A senior software developer employed for more than 11 years by Fujifilm Data Management Solutions in Australia was dismissed for serious misconduct on 7 August 2025 after a workplace bullying complaint escalated into a stream of complaint emails to managers. In [2026] FWC 835 (18 March 2026) the Fair Work Commission recorded that he agreed he relied on an AI large language model to prepare the emails, found that drafts were generated by AI, and recorded managers' unchallenged evidence that he appeared to consult ChatGPT during meetings. It found the emails inappropriate, combative and disproportionate, and said his use of AI 'was counterproductive. It led to his demise', appearing to give him a false sense of security. Because he accepted responsibility, the tribunal did not treat the AI use as an excuse; it held the dismissal was not unfair and dismissed his application. A Full Bench refused permission to appeal on 2 July 2026.

AI system
Unidentified chatbot
Occurred
Jun 2025 to Aug 2025
Reported
18 March 2026
Event location
Australia
What the AI did
Communicated with the person
Reported harm
Professional HarmLegal Harm
Whose AI use
Their own AI use
Setting
Work · Justice
Evidence
AI involvement supported · Causal attribution supported · 3 sources, 2 underlying accounts
6 claims: 6 documented. 2 open questions
People reported harmed
1 person

AI system as recorded: An AI large language model the employee used to draft workplace complaint emails, with managers' unchallenged evidence that he appeared to consult ChatGPT during meetings, per the Fair Work Commission

What Happened

Background. The applicant worked for Fujifilm Data Management Solutions as a Senior Java Developer from October 2013 until 7 August 2025, with no prior complaint about his performance or conduct, according to the Fair Work Commission. A disagreement with a colleague on 30 May 2025 led to a bullying complaint and then to further complaints and an extended exchange of emails with several managers.

The AI use. The Commission recorded that he agreed he relied on 'an artificial intelligence large language model (AI)' in preparing the emails, and found that drafts of the emails 'were generated by AI'. Managers' unchallenged evidence was that he 'appeared to be referring to Chat GPT for guidance' during meetings. He described his use of AI as 'a process of collaboration' and accepted ultimate responsibility for the communications.

The tribunal's findings. The Commission found the emails inappropriate, 'disproportionate to the matters being dealt with' and lacking nuance. On AI, it said his use 'was unfortunate. It was counterproductive. It led to his demise', that the communications it produced were 'demanding and overbearing', and that the AI use appeared to give him 'a false sense of security' that they were acceptable. Because he took responsibility for the communications, it did not regard the AI use as an explanation for his conduct.

Outcome. Accepting the employer's description that he had become 'ungovernable', the Commission held on 18 March 2026 that the dismissal was not harsh, unjust or unreasonable. It weighed his long service and his statement that the dismissal had forced him to retire as factors in his favour, but found them outweighed by his misconduct. A Full Bench refused permission to appeal on 2 July 2026. The Commission also said he 'was not wise' to prepare his tribunal case with AI.

Reported harm

The Fair Work Commission upheld the developer's dismissal and dismissed his claim, finding his reliance on AI-drafted communications counterproductive and that it led to his dismissal; he told the tribunal the dismissal had forced him to retire.

How Harm Occurred

An employee relied on a large language model to draft a stream of workplace complaint emails; the tribunal found the AI-produced communications combative and disproportionate, that the AI use gave him a false sense of security, and that it led to the dismissal it upheld.

Outcome

Resolved

The Fair Work Commission found the dismissal was not harsh, unjust or unreasonable and dismissed the unfair dismissal application on 18 March 2026 ([2026] FWC 835). A Full Bench refused permission to appeal on 2 July 2026 ([2026] FWCFB 159). The employer paid five weeks' pay in lieu of notice according to the termination letter quoted in the decision.

What remains unknown

  • Which AI product or products the employee used to draft the emails; the managers only believed it appeared to be ChatGPT.
  • Which of the listed emails were AI-drafted and to what extent he edited them.

What the evidence supports

AI involvement: supported. The Fair Work Commission found that drafts of the employee's complaint emails were generated by an AI large language model he relied on, which he did not deny, and recorded unchallenged evidence that he appeared to consult ChatGPT in meetings. It found the AI-produced communications 'dense, repetitive and often rambling' and 'demanding and overbearing', and that his AI use was counterproductive, gave him a false sense of security and 'led to his demise'. The dismissal it upheld rested on conduct that included those communications.

6 claims: 6 documented. What the statuses mean

Documented According to the Fair Work Commission's decision, the applicant was employed by Fujifilm Data Management Solutions as a Senior Java Developer from 8 October 2013 to 7 August 2025 and was dismissed for serious misconduct.

Causal attribution. The decision's statement of undisputed background.

  • fwc.gov.au(opens in new tab) supports · English
    'from 8 October 2013 to 7 August 2025. He was employed as a Senior Java Developer. He was dismissed for serious misconduct.'
Documented The Commission recorded that the applicant agreed he relied on an artificial intelligence large language model in preparing his many emails to managers, found that drafts of the emails were generated by AI, and recorded unchallenged evidence from managers that he appeared to consult ChatGPT on his phone during meetings; the applicant described his use of AI as collaboration and accepted responsibility for the communications.

Causal attribution. The tribunal's own findings and its record of the applicant's admission; HCAMag relays the same decision.

  • fwc.gov.au(opens in new tab) supports · English
    'he relied upon an artificial intelligence large language model (AI)'; 'emails were generated by AI.'; 'appeared to be referring to Chat GPT for guidance when responding to management comments'; 'He described his use of AI as a process of collaboration.'; 'He accepted ultimate responsibility for what was contained in all communications.'
  • hcamag.com(opens in new tab) supports · English
    'He later admitted drafting them with AI assistance, and colleagues observed him consulting what appeared to be ChatGPT during meetings.'
Documented The Commission found the emails inappropriate: often needlessly detailed and written in a combative and demanding tone, disproportionate to the matters being dealt with, lacking regard to purpose or context, and lacking nuance.

Causal attribution. Findings of the tribunal.

  • fwc.gov.au(opens in new tab) supports · English
    'The signs of this included that they were often needlessly detailed, many were written in a combative and demanding tone that was inappropriate for communications in a workplace setting'; 'The emails were inappropriate. They were disproportionate to the matters being dealt with. They lacked any reasonable regard to purpose or context. They lacked nuance.'
Documented The Commission found the applicant's use of AI 'counterproductive' and that it 'led to his demise', that the communications it produced were 'dense, repetitive and often rambling' and 'demanding and overbearing', and that the AI use appeared to give him a false sense of security that the communications were acceptable; because he took responsibility for them, it did not treat the AI use as an explanation for his conduct.

Causal attribution. The tribunal's causal assessment of the AI use, stated in its own words; it is the tribunal's finding, not an independent technical analysis.

  • fwc.gov.au(opens in new tab) supports · English
    'use of AI was unfortunate. It was counterproductive. It led to his demise.'; 'The communications it produced were dense, repetitive and often rambling. They were demanding and overbearing.'; 'a false sense of security that his communications, laden as they were with allegations of impropriety by his managers and demands for corrective action, were appropriate and acceptable in a workplace setting. Objectively, they were not.'; 'use of AI as an explanation for his conduct.'
Documented On 18 March 2026 the Commission found the dismissal was not harsh, unjust or unreasonable, accepting the employer's description that the applicant had become 'ungovernable', and dismissed his unfair dismissal application; it treated his long service and his statement that the dismissal had forced him to retire as factors in his favour.

Causal attribution. The decision's conclusion.

  • fwc.gov.au(opens in new tab) supports · English
    'SYDNEY, 18 MARCH 2026'; 'dismissal was not harsh, unjust or unreasonable'; 'had become ungovernable. His communications with his employer were inappropriate.'; 'and the dismissal has forced him to retire'
Documented On 2 July 2026 a Full Bench of the Commission refused permission to appeal against the decision.

Causal attribution. The appeal decision.

Sources

3 sources inspected, from 2 underlying accounts. Sources that repeat one account do not corroborate each other.

How the sources were read, and where the events happened

Read in English on 2026-10-21: the Fair Work Commission's published decision [2026] FWC 835 (Deputy President, Sydney, 18 March 2026), PDF from fwc.gov.au (HTTP 200), text extracted with pdftotext; whole decision available and the background, 'The use of AI' section and conclusion were read. Applies to s1.

Read in English on 2026-10-21: Full Bench decision [2026] FWCFB 159 (2 July 2026) refusing permission to appeal against [2026] FWC 835, PDF from fwc.gov.au (HTTP 200). Applies to s2.

Read in English on 2026-10-21 (HTTP 200). HRD Australia (HCAMag) report published 23 March 2026 summarising [2026] FWC 835; it relays the decision, so it shares the decision's independence group. Applies to s3.

Event countries: Australia. Affected-person countries: Unknown. Court countries: Australia.

The employer is an Australian company, the decision was made by the Fair Work Commission in Sydney under the Fair Work Act 2009, and the employment and dismissal were dealt with under Australian law. His residence is not stated.

Reviewed for publication 2026-10-21: Published on the Fair Work Commission's own findings: drafts of a dismissed developer's complaint emails were generated by AI, the AI use was counterproductive and 'led to his demise', and the dismissal was upheld, with permission to appeal refused. The applicant is not named in the record.

People described

A senior software developer in Australia, employed for more than 11 years, whose dismissal the Fair Work Commission upheld

People reported harmed in this case

1 person

1 AI participant · 0 other people harmed

One person: the dismissed developer who used the AI tool, per the Fair Work Commission's decision.

Counted once within this case. The same person may appear in other cases. This count does not establish AI causation.

Tags

#employment#dismissal#fair-work-commission#ai-drafted-emails#consequential-response#tribunal-finding

Cite this case

Compiled per our published methodology: verification statuses, sourcing standards, and corrections process.

APA

NOPE. (2026). Fair Work Commission in Sydney upholds a software developer's dismissal, finding his reliance on AI-drafted complaint emails was counterproductive and led to it. AI incidents. https://nope.net/incidents/2025-sydney-fair-work-commission-upholds-software-developers-dismissal-finding-reliance-on-ai-drafted-emails-led-to-it

BibTeX

@misc{2025_sydney_fair_work_commission_upholds_software_developers_dismissal_finding_reliance_on_ai_drafted_emails_led_to_it,
  title = {Fair Work Commission in Sydney upholds a software developer's dismissal, finding his reliance on AI-drafted complaint emails was counterproductive and led to it},
  author = {NOPE},
  year = {2026},
  howpublished = {AI incidents},
  url = {https://nope.net/incidents/2025-sydney-fair-work-commission-upholds-software-developers-dismissal-finding-reliance-on-ai-drafted-emails-led-to-it}
}

If you or someone you know is struggling, free and confidential support is available. Find a helpline near you at Signpost.

Back to all incidents