In a striking clash between artificial intelligence and legal accountability, a Texas judge has dismissed a defense by Autofit, an aftermarket auto parts company, which claimed that a ChatGPT-generated document was responsible for an unlawful termination confession. The case, which has caught the attention of labor law experts, underscores the complexities of AI use in corporate settings.

Daniela Melendez, a former administrative assistant at Autofit, was terminated just sixteen days into her job after discussing her salary with a colleague. When Melendez applied for unemployment benefits, Autofit, through an administrative assistant, submitted a written statement to the Texas Workforce Commission explaining the termination due to "performance issues" and "breaches of confidentiality," specifically citing the sharing of pay details among colleagues.

However, discussing pay is not illegal, whereas firing someone for this reason is against labor laws. When pressed for clarification, Autofit's initial attempt to reinterpret the statement did not succeed, leading to an extraordinary second defense during trial: blaming the AI tool ChatGPT for the inappropriate content of the statement.

Autofit’s attorney argued that the statement was composed using the free version of ChatGPT to craft a professional-sounding response, which was then directly pasted into the official form. The employee responsible admitted to not thoroughly reviewing the document, assuming any errors would be corrected by the automated document processing service ADP.

Judge Sharon Steckler, presiding over the case, found this defense implausible. She noted that even if ChatGPT had generated the content, it was Autofit's responsibility to review and correct any inaccuracies before submission. The judge emphasized, "Guevara would have us believe that ChatGPT made up the language that sensitive information includes pay... I therefore do not credit Guevara’s explanation that ChatGPT added pay without Guevara’s help."

Further complicating Autofit’s position, the company presented numerous affirmative defenses, none of which prevailed. The judge's ruling favored Melendez, ordering her reinstatement, back pay, compensation for pecuniary harms, expungement of her record, and the posting of a notice detailing the case outcome for sixty days at the company's premises.

The case raises significant questions about the reliance on AI tools in sensitive corporate processes and the ethical implications of such use. As machines play an increasingly prominent role in work environments, the Autofit case serves as a cautionary tale about the potential legal and moral hazards of substituting human oversight with artificial intelligence.