For a year and a half, Yves Valerus had steady full‑time work as a Haitian Creole–English interpreter: set hours, an hourly rate and benefits. She handled hospital, court and other vital calls by phone and video. In 2025, her employer said business had slowed and introduced new scheduling software. Her shifts became fragmented and unpredictable; by the end of the year her pay was nearly 20% lower than the previous year. As a single mother in Brooklyn raising three children, Valerus scrambled—prioritizing internet service to keep working remotely and changing how she shopped to stretch the reduced income.
Valerus and some colleagues have petitioned to organize with the Communications Workers of America and worry the company may be piloting AI to handle basic interpretation. Their experience reflects a broader trend: algorithmic scheduling and other management technologies that shift business risk onto hourly workers and intensify on‑the‑job demands.
The employer is LanguageLine Solutions, owned by call‑center giant Teleperformance, a firm that has previously faced accusations about worker surveillance and has reached union agreements over monitoring. Interpreters describe emotionally demanding work that ranges from explaining insurance coverage to translating a judge’s sentence. They say ordinary downtime has largely disappeared: what used to be a minute or two between calls is often reduced to a mandatory 15 seconds. The compressed pace contributes to fatigue and mistakes; video interpreters say they barely have time to stand or stretch.
LanguageLine responded that it “absolutely and categorically” does not want unsafe conditions and that health and safety committees review workplace issues. The company also said it faces industry‑wide headwinds, adjusted schedules because volume was lower than expected, and is piloting AI for routine tasks it believes could expand the market rather than eliminate jobs.
Algorithmic tools are not unique to call centers. Hotel cleaners, for example, have resisted software that sequences rooms in ways that increased injury risk by forcing heavier lifting and unnecessary running between floors. “These are not pre‑ordained outcomes of the software,” said Carlos Aramayo, president of UNITE HERE Local 26. “They are management decisions.”
At LanguageLine, one concrete scheduling change showed up as unpaid “mandatory involuntary time off,” coded AEX. Interpreters could be fully scheduled one day and see AEX on the next day with little or no advance notice—even learning about the change midshift. That unpredictability makes it difficult to take other gigs or plan daily life. Tax documents reviewed by NPR showed Valerus’s income fell 18% from 2024 to 2025; another part‑time worker saw pay drop by more than 70%. Interpreter Anna Manciano said she left the job after having a baby because income was too unstable. Workers also say a company code of conduct barred them from freelancing for other translation services while their hours were being cut.
The timing of these disruptions coincided with LanguageLine’s adoption of NiCE, a workforce management platform that markets “smarter scheduling, accurate forecasting, and real‑time intraday optimization,” including AI‑driven features. NiCE’s marketing emphasizes labor‑cost reductions, and it lists major corporate clients. LanguageLine said NiCE was used to forecast volumes and set schedules and that it factors in real‑time events such as weather and travel disruptions. NiCE did not respond to requests for comment.
Researchers say detailed before‑and‑after stories matter. Daniel Schneider, a lead on Harvard’s SHIFT project, noted that algorithmic scheduling became common in retail over a decade ago and helped spur “fair workweek” laws in some places; those protections, however, are mostly limited to retail, leaving many sectors unprotected. Susan Lambert of the University of Chicago and Virginia Goellgast, who studies unionized call centers at Cornell, argue that unionized workplaces tend to fare better because unions can negotiate scheduling rules and protect workers from the economic fallout of algorithmic management.
Workers at LanguageLine pushed back in 2025: more than 200 interpreters signed a petition protesting reduced hours. Since 2024 some employees have been organizing for a union and are calling for higher pay, paid bathroom breaks and stronger benefits. The campaign has drawn political attention; at a press event at New York City Hall, Comptroller Mark Levine urged LanguageLine to respect workers’ union rights and said the city was reviewing its contracts with the company for compliance.
LanguageLine has acknowledged that schedules have been “more unpredictable than normal” and said it is calibrating systems to improve stability. Workers, however, say the software has shifted risk away from the company and onto employees by compressing downtime and pay. Valerus remains active in organizing while looking for additional work to support her family. “This is very stressful,” she said.