We had a bit of an AI surprise this week.
For one of our clients, we produced a series of videos featuring people sharing very personal stories about their mental health. One participant told her story in Spanish.
Then she reached out to us with a concern: When she watched the video in English on YouTube, it wasn’t her voice telling her story. It was someone else’s.
We were confused. Really confused. We hadn’t produced an English version.
After some digging, we figured it out: YouTube had automatically created an English dub of her video using AI.
And on one hand, that’s pretty cool. Someone who doesn’t speak Spanish can now hear and understand a story they otherwise might not have been able to. That’s exactly the kind of thing AI can be really good at: making information more accessible and removing barriers.
But there was a pretty big problem.
This wasn’t just information. It was her story.
She had trusted us to help her share something deeply personal, in her own words and her own voice. Hearing an AI-generated voice suddenly telling that story understandably felt very different.
We turned the feature off and addressed the issue. But it was a good reminder for me about the way AI is increasingly showing up in our work.
We spend a lot of time talking about whether we should use AI. But increasingly, AI is just there. It’s being built into the platforms and tools we already use. Sometimes turned on by default, and sometimes it does things we didn’t explicitly ask it to do.
And while those features are usually designed to help, “helpful” isn’t the same thing as “right for every situation.”
A tool that makes a video more accessible can also make someone’s story feel less like their own. A tool that saves us time can remove a moment when a human might have stopped and asked, Should we actually do this?
I’m not anti-AI. Far from it. There are plenty of ways it can make our work better.
But this was a good reminder that the more AI can do for us, the more important it becomes to pay attention to what it’s doing on our behalf.
And one practical takeaway: If you manage a YouTube channel, it’s worth checking your automatic dubbing settings. You may have AI-generated versions of your videos that you didn’t intentionally create. Depending on the content, especially when real people are sharing personal or sensitive stories, that could do more than create an awkward surprise. It could unintentionally cause harm.
AI is going to keep showing up in places we don’t expect it. We just need to make sure we’re still paying attention.
