Royal Mail has confirmed that it used an artificial intelligence-generated voiceover in its 2025 Christmas advertising campaign after the daughter of late British broadcaster Paul Vaughan claimed the narration sounded remarkably similar to her father’s voice. The revelation has sparked concerns about the potential reproduction of deceased performers’ voices without permission and the growing use of synthetic voices in commercial advertising.
Vaughan, who died in 2014 at the age of 89, was widely recognized for narrating the BBC science documentary series Horizon, BBC Radio 4’s Kaleidoscope, and the Paddington audiobooks. His daughter, Lucy Wynne, said she was shocked when she heard what she believed was her father’s distinctive voice in a Royal Mail television advertisement in November 2025, more than a decade after his death.
Royal Mail Confirms AI-Generated Narration
The advertisement, part of Royal Mail’s Christmas campaign promoting personal connections through greeting cards, featured real actors but used an AI-generated narrator. The campaign included a closing message emphasizing that sending a card represents more than a simple gesture.
Wynne said she immediately recognized similarities to her father’s voice while watching the commercial at home. Her husband also believed the narration sounded like Vaughan, prompting concerns about whether recordings of his previous work had been used to create the synthetic voice.
Royal Mail subsequently acknowledged that the voiceover was generated using artificial intelligence but denied that the production had intentionally recreated Vaughan’s voice. The company explained that external production partners had developed the narration using prompts rather than attempting to imitate a particular individual. It also said its partners had conducted checks that did not identify concerns about the voice resembling an existing performer.
According to Deadline, the advertising campaign involved agency Darwin & the Machines and production company Untold Studios, while the AI narration was handled by Czech company Fameplay.
Vaughan’s Daughter Questions the Use of His Voice
Wynne has strongly criticized the advertisement, arguing that neither she nor her family had authorized the use of her father’s voice. She described the experience of unexpectedly hearing a voice resembling his on television as deeply upsetting, particularly because Vaughan had been dead for 11 years when the commercial aired.
The controversy has also attracted attention from voice professionals and researchers. Professor Vincent Hughes, a forensic speech scientist at the University of York, examined the advertisement’s narration alongside recordings of Vaughan and identified considerable phonetic similarities between the voices. However, the analysis did not establish that Vaughan’s recordings had been used to generate the advertisement.
The AI technology company ElevenLabs, whose technology was reportedly licensed for the production, denied that the narration was a direct clone of Vaughan. The company did not confirm whether recordings of his previous performances had been included in the training data used by its technology.
ElevenLabs has since added Vaughan to its No Go Voices list, which is intended to prevent direct cloning of protected voices through its platform.
Voice Actors Raise Concerns Over AI Advertising
The incident has renewed discussion about consent and the use of artificial intelligence in voiceover production. Peter Caulfield, an actor associated with the Save Our Voices Now campaign, criticized the risks of using synthetic voices without sufficient safeguards, particularly when the resulting audio resembles a recognizable performer.
Wynne also questioned the decision to use AI narration instead of employing a professional voice actor, saying her father would have opposed technology being used to replace working performers.
Although Royal Mail has acknowledged using AI, there has been no confirmed finding that Vaughan’s voice was deliberately cloned or that his recordings were used without authorization. The dispute nevertheless highlights the difficulties facing performers and their families when synthetic narration closely resembles a recognizable human voice.
The controversy comes as voice actors continue to seek stronger protections governing the use of their recorded performances in AI systems, particularly when it comes to consent, compensation, and the reproduction of voices after death.