AI announcer mispronounces, skips names at glendale community college graduation

An AI-powered announcer mispronounced students’ names and skipped others entirely during Glendale Community College’s commencement in Phoenix, Arizona, forcing the ceremony to pause at least twice [The Verge]. College president Tiffany Hernandez apologized and offered affected students a chance to re-walk the stage.

The system was meant to ensure accurate name pronunciation—a core function it failed to deliver. Instead, it stumbled on phonetics and timing, cutting off or omitting names as graduates approached the stage. One student reported hearing their name garbled into an unrecognizable sequence; others weren’t called at all.

This wasn’t a minor glitch—it disrupted a milestone event where correct name announcement is both symbolic and personal. The errors reflect broader issues in speech synthesis: even with pre-loaded name data, AI struggles with uncommon or non-English phonetic structures. Timing misfires suggest poor integration with live event pacing.

The incident underscores a recurring pattern: AI deployments in emotionally significant, real-time settings often lack sufficient fail-safes. Here, human moderators had to intervene repeatedly, exposing the fragility of automated systems when stakes are high.

There’s no indication the college tested the system with a diverse sample of names under live conditions. If they did, the flaws were ignored. Either way, the rollout reflects overconfidence in unproven automation.

No student should have to reapply for a moment that should happen once, correctly. When AI disrupts rather than supports, it’s not innovation—it’s negligence.