Can a Friendly Voice Make a Realistic AI Presenter Less Creepy?

An artificial intelligence presenter can have convincing facial features, realistic expressions and an almost human appearance, yet watching it may still feel strangely uncomfortable. New research suggests that the sound of a digital presenter's voice could help explain why some AI characters are more appealing than others.

Researchers investigating AI-generated livestream presenters found that highly realistic characters could provoke stronger negative emotional reactions than less lifelike alternatives. However, a warmer voice helped reduce those responses, offering an interesting insight into how audiences experience artificial personalities.

Why do realistic AI faces sometimes feel unsettling?

The phenomenon is known as the uncanny valley, a term describing the discomfort people can experience when an artificial character looks almost human but still appears slightly unnatural.

A cartoon presenter may be easy to accept because viewers immediately recognise it as artificial. A photorealistic character creates different expectations, particularly around facial movements, expressions and speech.

Researchers Crystal T. Lee and Ling-Yen Pan investigated these reactions through four experiments involving AI-enabled livestream presenters.

Their findings indicated that highly realistic digital characters could produce stronger negative emotional responses, which were associated with reduced audience engagement and greater avoidance.

One experiment also used facial-analysis technology to examine participants' reactions while they watched the presenters.

What difference does a friendly voice make?

The researchers explored whether vocal warmth could influence the way viewers responded to realistic AI characters.

A warmer voice helped reduce the negative effects associated with highly lifelike presenters, suggesting that audiences judge digital personalities through a combination of appearance and sound.

Another experiment examined six levels of visual realism. When vocal warmth was lower, audience responses followed a pattern consistent with the uncanny valley effect, while warmer voices reduced that pattern.

The findings do not mean that any realistic AI character will become appealing simply by using a pleasant voice. They suggest that developers should consider how different features work together rather than concentrating entirely on visual detail.

Could this influence television and entertainment?

Digital presenters are being explored for online video, livestreaming and commercial communication, where they can deliver information and interact with audiences without conventional filming arrangements.

The study focused on AI-enabled livestreaming rather than British television broadcasting. It did not establish whether audiences would prefer digital newsreaders or presenters to human performers.

Nevertheless, the findings could be relevant to entertainment producers developing artificial characters. A presenter with a stylised appearance, natural speech and expressive movement may be more engaging than a highly realistic character whose behaviour feels unnatural.

The research suggests that increasing realism is not necessarily the same as improving the viewing experience.

Will AI presenters become more convincing?

As AI technology develops, digital characters are likely to become increasingly sophisticated in their appearance, movement and speech.

However, creating a technically impressive presenter is only part of the challenge. Producers also need to understand whether audiences enjoy watching the character and feel comfortable interacting with it.

The study highlights why human reactions should remain central to the design process. Rather than simply making artificial presenters look more like people, developers may achieve better results by considering how they sound, behave and communicate.

Research and sources: Crystal T. Lee and Ling-Yen Pan, International Journal of Information Management, October 2026. Read the original study.