Why do some followers think our social videos are voiced by AI?

AI technology has improved so much that synthetic voices can sound surprisingly human, reports ITV News Digital Video Producer Isaine Blatry

A few months ago, I started noticing an interesting trend in the comments section of our TikTok videos.

People were questioning whether my voice was real or AI-generated. It took me by surprise because, well, I am real!

As someone who speaks English as a second language, I know I have an accent, which got me thinking: Why are more and more people mistaking my voice for AI?

Wanting to understand what might be driving this perception, I decided to dig deeper into the world of AI-generated voices.

AI-generated voices are everywhere now, from social media platforms to voice assistants like Siri and Alexa. We hear them in customer service calls, ads, and even video narration.

But in the early days, these voices were robotic, monotone, and easy to distinguish from human speech.

Today, things have changed. AI technology has improved so much that synthetic voices can sound surprisingly human.

"Typically, AI deep fakes or audio deep fakes will sound quite monotone," Dr Michael Cook, a professor in Computer Science at King’s College London tells me.

"They don't have the same kind of emotion that you might expect from someone speaking naturally."

Even so, many modern AI voices have become far more convincing than they used to be, which may explain why some people question the authenticity of real voices like mine.

One of the reasons people might think my voice is AI is because of how it sounds when I speak English, my second language.

Being French, I naturally have an accent when speaking English, and accents can sometimes throw people off.

As Dr Cook explains: "One factor, for example, is that you're a young woman, which many voices are trying to sound like. Another factor could be that you're speaking in a way that is less familiar to people."

But it’s not just accents.

Al Black, the director of Voquent - a voice actor’s agency - told me another layer to this issue: “[Artificial intelligence[ voices sound perfect and flawless. I think it's easy to forget that humans are organic creatures - we breathe.”

In other words, when people listen to heavily edited audio in videos, where breaths and background noise are cleaned up, the voice can sound too perfect, which might be another reason why people think it’s AI.

In my investigation, I found out that one of the reasons AI voices can feel so detached is that they lack the human touch - especially the kind of emotional nuance that real people bring to their speech.

Margarita Grubina, who works with AI-generated voice technologies at Respeecher, explains it perfectly, telling me: “AI is still not there to replace humans in terms of emotions and performance.

"With AI, you can pick if you want it to sound sad or happy or angry, but it still doesn’t reach the complexity of human emotions.”

When people mistake my voice for AI, I wonder if it’s because AI voices often try to sound neutral and without any regional accent.

Many of the voiceovers we hear from AI have this flawless quality that erases things like accents or emotional variance. This might make them sound unnatural but also hard to distinguish from highly processed human speech.

AI-generated voices are becoming more common in social media videos, too.

Some companies are even moving entirely to AI for their content, paying real people for the rights to use their voice but adapting it to say whatever they need like Respeecher.

This raises a new question: Could AI one day replace human content creators altogether?

Subscribe free to our weekly newsletter for exclusive and original coverage from ITV News. Direct to your inbox every Friday morning.

For now, at least, the answer seems to be no.

Mr Black points out that AI voices still don’t resonate with audiences in the same way real voices do: “It gets very tiring on the ears to listen to somebody who speaks at the same tone and pitch constantly.”

And that’s where the real difference lies. AI can deliver facts, but it can’t connect with us emotionally. Social media, especially platforms like TikTok, thrives on emotional connection.

Whether it’s the enthusiasm in someone’s voice or the subtle emotional undertones, these are the things that AI can’t yet replicate.

So why do some people think my voice is AI?

It could be a mix of factors: my accent, the processing of my audio, and the fact that AI voices have become so advanced that it’s sometimes hard to tell the difference between human and machine.

But as I’ve learned through my research, there’s still something special about a real human voice that AI can’t capture—something authentic and emotionally resonant.

As AI technology continues to evolve, it will become even more difficult to distinguish between synthetic and human voices.

But I believe there’s always going to be a place for real, unfiltered human speech in the digital world. Because at the end of the day, we connect more deeply with voices that breathe, pause, and carry the weight of real emotions.

And no matter how advanced AI becomes, that’s something it will never truly replicate.

Want a quick and expert briefing on the biggest news stories? Listen to our latest podcasts to find out What You Need To Know...