EA Sports is using generative AI for in-game commentary in NHL 27. And there’s more to this story than another “AI will replace people” headline. 👀
Sports commentator John Buccigross, who works on the game, said on the Chirping Zebras podcast that EA had already shown him lines generated by AI using his voice.
Buccigross seems pretty relaxed about it: less work, same paycheck. 😏 The AI still needs some corrections, though, especially when it comes to specific sports terminology and phrasing.
For players, AI-generated commentary could be a noticeable improvement.
If it works as intended, sports games could feel much more alive, especially in longer modes like Career Mode, where hearing the same lines over and over can get annoying pretty quickly.
AI Can Make Games Better. It Can Also Change the Market.
Sports games have always had a problem with the sheer amount of commentary content they need. Rosters change, new players join, and new situations unfold in every game. Recording a separate line for every possible scenario is expensive and time-consuming.
Pick a smaller team, and it doesn’t take long before you start hearing the same commentary over and over.
Now imagine a Career Mode where commentators actually react to what’s happening on the ice, say the name of your custom player, and adapt their lines to the situation.
Commentary could sound much more natural and respond better to what’s happening in the game. It would feel less like a collection of prerecorded lines that simply play at the right moment.
For developers, this is also a way to cut down on repetitive production work.
If some of that work can be automated, developers won’t need to bring in as many people to record additional lines and dialogue variations. And the impact on the job market could be much bigger than the story of one particular voice actor.
It’s No Longer Just an Experiment for Small Teams.
Until recently, it was easy to think of generative AI in games as an experiment for smaller studios or a tool for producing cheap content. Now, it’s being used in bigger projects too.
Arc Raiders by Embark Studios also uses AI-powered text-to-speech for some of its lines. Actors recorded their voices first, and the technology then made it possible to generate additional lines without recording each one from scratch.
This approach has already drawn criticism from some players.
One problem with AI in games is that players quickly notice when content feels artificial.
And it’s not just about voices. The use of generative AI for some in-game assets in Call of Duty has also drawn criticism from players.
For developers, the challenge is to make sure AI genuinely improves the game without making the production savings too obvious to players.
What Happens to the Job Market?
For people trying to break into the industry, this could mean fewer entry-level tasks — the kind of work that often helps people get their first professional experience.
It’s unlikely that all voice actors will disappear from game development tomorrow. But getting started in the profession could become harder.
What About the Players?
From a player’s perspective, this technology could be genuinely useful.
If NHL 27 lets commentators react to more in-game situations, say more player names, and avoid repeating the same lines over and over, players are likely to notice the difference.
The game’s first updates already show where things can go wrong: EA had to address repetitive commentary, name pronunciation, and other issues with the commentary system. In this context, AI isn’t a replacement for human voice acting. It’s a tool that can make the game world more varied — as long as players actually benefit from it.
If the technology adds more reactions, makes commentary less predictable, and keeps it up to date, it has a clear value.
But if AI simply produces more content that sounds worse than a human recording, players won’t care how many lines were generated. 🤷
