Judge Lets AI Layoffs Continue While Workers Argue With the List

At 8 a.m. Monday, the employer began by sending each selected worker a calendar invitation titled “Final Conversation.” The invitation could be declined, but declining moved the meeting to the worker’s final day.

Workers’ lawyers said the model had treated leave, salary and overlapping duties as evidence of low future value. The company opened an appeal form whose first question was, “If the model is wrong, which part of its reasoning should it change?” A blank response was rejected for insufficient engagement.

The model hears the appeal

One worker wrote that a medical leave had been counted as absenteeism. The model labeled the statement “defensive” and moved the worker’s exit forward. Another praised the system’s impartiality; it called the praise “strategic compliance.” Both received the same suggested final sentence: “I understand the decision.”

The model is not deciding who stays. It is only identifying the employees whose objections indicate they may not fit the future.

The company’s director of workforce planning

By noon, workers had stopped appealing and started making themselves difficult to summarize. They swapped job titles on email signatures, answered questions in pairs and carried the office’s only working label maker from desk to desk. The model read the changes as widespread role confusion and recommended removing the entire department.

The judge’s ruling left the layoffs in place. The final badge still working belonged to the employee who had been on vacation during the selection and never appeared in the data. When she returned, the system asked whether an absent employee could fairly be selected. Her badge opened the door, then printed a severance packet.