AI System Reads Names Wrong at Arizona College Graduation, College Has to Fix It on the Spot

An artificial intelligence system that was supposed to read graduate names during Glendale Community College's graduation ceremony on May 15 started skipping names and saying them incorrectly. The college's leadership had to step in and fix the problem mid-ceremony after the crowd reacted negatively.

What Went Wrong

The college was using a new AI system designed to announce each graduate's name. But the system made mistakes: it left some names out completely and mispronounced others. KSDK reported that the college described this as a "technical issue" with their new system.

College President Tiffany Hernandez took the stage to explain what had happened. Yahoo News reported that she told the crowd the problems came from the "new AI name-reading system they were using."

It appears the college tried this system for the first time without having a backup plan if things went wrong. The AI had trouble reading the list of graduate names correctly, which suggests the underlying software or the way the names were prepared for the system had problems.

How the College Responded

At first, President Hernandez told the graduates whose names had been skipped or mispronounced that they would not get a second chance to walk across the stage and have their names announced properly.

The crowd didn't like that decision. There was audible booing from the audience.

The college quickly changed its mind. The graduates whose names had been missed or gotten wrong were called back up, and everyone got their moment recognized. The ceremony finished with all graduates properly celebrated.

Why This Matters

Graduation ceremonies are a tough place for a computer system to fail. Unlike other things that go wrong at work or school and can be fixed quietly later, a graduation ceremony happens once and is full of emotion for families and students. If something breaks, there is no real way to fix it afterward.

The AI system the college used is called text-to-speech. It converts a written list of names into spoken words. This type of system works pretty well most of the time, but it struggles with certain names — especially names from languages other than English. Names with unusual letters or sounds are like the tricky parts of the map for a GPS system: the system doesn't know what to do with them.

The college serves a diverse group of students, which means the system was trying to handle a wide range of name types. This made the job harder for the AI, which may have been trained mainly on English names.

The broader context here is that many organizations are eager to use AI to do new things and look modern, but they're deploying these systems before they really know if they work in real situations. Graduation ceremonies are high-stakes moments where mistakes are noticed immediately and cannot be undone. This is not the right place to test new technology.

What Should Have Been Different

A well-designed system would have had several checkpoints to catch problems. Before the ceremony, someone should have checked the entire list of names to make sure the AI could handle them. During the ceremony, a person should have been monitoring the system and ready to take over if something went wrong. And the college should have trained staff to switch back to having a person announce the names if needed.

The college also didn't think through what would happen if the AI failed. A good plan would have said: "If the system breaks down, here's what we do." Instead, the college made decisions about what to do after things had already fallen apart.

Moving Forward

Other colleges and schools will probably learn from what happened at Glendale. This incident shows that even simple-sounding uses of AI—like reading names out loud—actually need careful planning, testing with real ceremonies, and real people ready to step in when technology fails.

The technology itself isn't broken. Text-to-speech systems work quite well in controlled situations. But using them in front of hundreds of people at an important moment requires a different kind of thinking: you need to be ready for what goes wrong, not just hope that it doesn't.