British voice actor says an AI clone of his voice, built from an old IBM recording and sold to websites, has cost him jobs
Greg Marston, a British voice actor, said in accounts published in 2023 that an AI-generated clone of his voice, built from a recording he had made for IBM years earlier, was being sold to websites that were using it to build a synthetic voice that could say anything. Julia Angwin's New York Times column (republished by the Oman Observer on 25 September 2023) says he found the clone, called Connor, online and heard his voice on the Wimbledon website during the tennis tournament; Cryptopolitan (1 July 2023) says he found it on Revoicer, an AI text-to-speech site. He had signed a release allowing the recording to be used in many ways, but said he never agreed to having his voice cloned to compete against him. He estimated that the clone had already lost him jobs and was working with a lawyer to seek compensation. IBM said it was aware of his concern and was discussing it with him. The outcome is not reported in the inspected sources.
- AI system
- Revoicer text-to-speech (reported)
- Occurred
- Event date unknown
- Reported
- 1 July 2023
- Event location
- Unknown
- What the AI did
- Depicted or impersonated the person
- Reported harm
- Exploitation or AbuseProfessional Harm
- Whose AI use
- Someone else’s AI use
- Setting
- Work · Privacy
- Evidence
- AI involvement reported · Causal attribution alleged · 2 sources, 1 underlying account
- 5 claims: 5 reported. 4 open questions
- People reported harmed
- 1 person
AI system as recorded: Synthetic AI voice, called Connor in the NYT column, built from a recording the actor made for IBM and sold to websites, including (per Cryptopolitan) the Revoicer text-to-speech site, to build a synthetic voice that could say anything
What Happened
The clone. According to Julia Angwin's New York Times column, Marston 'recently came across “Connor” online — an AI-generated clone of his voice, trained on a recording Marston had made in 2003', saying things he had never said. Cryptopolitan says he discovered his voice in use without his knowledge or consent on Revoicer, a website that uses AI to convert text into speech, and dates the recording to 2005, for a satnav system.
The release. The column says he had recorded a session for IBM and later signed a release allowing the recording to be used in many ways, and that IBM was able to sell the sample to websites using it to build a synthetic voice that could say anything. Cryptopolitan says the contract's fine print allowed IBM to sell his voice to third parties.
What he reports. He estimated that the clone had already lost him jobs and would cut into his future earnings, and was working with a lawyer to seek compensation. 'I never agreed or consented to having my voice cloned, to see/hear it released to the public, thus competing against myself,' he told the columnist. He also heard his voice on the Wimbledon website during the tennis tournament.
Responses and limits. IBM said it was aware of his concern and was discussing it with him directly. The year and purpose of the original recording differ between the sources, the Financial Times interview that other outlets cite was not read, and no result of his compensation claim is reported in the inspected sources.
Reported harm
A voice actor says an AI clone of his voice, built from an old IBM recording and sold to websites, was used without his consent and has already lost him jobs; he was seeking compensation (his account in two 2023 reports).
Outcome
UnknownIn 2023 he was working with a lawyer to seek compensation and IBM said it was discussing his concern with him directly. No result is reported in the inspected sources.
What remains unknown
- When he found the clone; the earliest dated report is 1 July 2023.
- The year and purpose of the original IBM recording (2003 in the column; 2005, for a satnav system, in Cryptopolitan).
- Which websites sold or used the clone beyond Revoicer and the Wimbledon website, and how many jobs he lost.
- The result of his compensation claim and of his discussions with IBM.
What the evidence supports
AI involvement: reported. The column describes an AI-generated clone of the actor's voice, trained on his IBM recording, that says things he never said and was sold to websites building a synthetic voice that could say anything; Cryptopolitan names Revoicer, an AI text-to-speech site. He connects the clone's use to lost jobs. The clone's origin rests on his account and the outlets' reporting; IBM said only that it was discussing his concern, and Revoicer has not been heard.
5 claims: 5 reported. What the statuses mean
Reported Julia Angwin's New York Times column reports that Greg Marston, a British voice actor, came across 'Connor', an AI-generated clone of his voice trained on a recording he made for IBM, saying things he had never said.
Causal attribution. The columnist's account based on her interview with the actor; the year of the recording differs between sources (2003 in the column, 2005 in Cryptopolitan).
- omanobserver.om(opens in new tab) supports · English
Greg Marston, a British voice actor, recently came across “Connor” online — an AI-generated clone of his voice, trained on a recording Marston had made in 2003. It was his voice uttering things he had never said.
Reported According to the column, he had signed a release allowing the IBM recording to be used in many ways, and IBM was able to sell the sample to websites using it to build a synthetic voice that could say anything.
Causal attribution. The column's account; no contract or IBM licence document was inspected.
- omanobserver.om(opens in new tab) supports · English
Back then, he had recorded a session for IBM and later signed a release form allowing the recording to be used in many ways.; IBM was able to sell Marston’s decades-old sample to websites that are using it to build a synthetic voice that could say anything.
Reported Cryptopolitan reports that he found his voice being used without his knowledge or consent on Revoicer, a website that uses AI to convert text into speech.
Causal attribution. Cryptopolitan's account, which does not cite its own interview; it is treated as the same witness's account.
- cryptopolitan.com(opens in new tab) supports · English
One day, he discovered his voice being used without his knowledge or consent, on a platform called Revoicer.; This website uses AI to convert text into speech in numerous languages and styles
Reported The column says he estimated that the AI clone had already lost him jobs and would cut into his future earnings, and that he was working with a lawyer to seek compensation; he said he never agreed or consented to having his voice cloned and released to the public to compete against him.
Causal attribution. The actor's own estimate and statement to the columnist; no lost booking or amount is specified.
- omanobserver.om(opens in new tab) supports · English
He estimates that his AI clone has already lost him jobs and will cut into his future earnings significantly. He is working with a lawyer to seek compensation. “I never agreed or consented to having my voice cloned, to see/hear it released to the public, thus competing against myself,” he told me.
Reported The column says he heard his voice on the Wimbledon website during the tennis tournament, and that IBM said it was aware of his concern and was discussing it with him directly.
Causal attribution. The column's account and IBM's statement as relayed in it.
- omanobserver.om(opens in new tab) supports · English
Marston recently discovered his voice emanating from the Wimbledon website during the tennis tournament. (IBM said it is aware of Marston’s concern and is discussing it with him directly.)
Sources
2 sources inspected, from 1 underlying account. Sources that repeat one account do not corroborate each other.
- Oman Observer (republishing The New York Times), 25 September 2023: The internet is about to get much worse(opens in new tab)
s1 · omanobserver.om · Opinion column · English · Inspected · 25 September 2023 · Shares an underlying account with another listed source · Primary
- Cryptopolitan, 1 July 2023: Epic disruption: AI robs actors of their voice(opens in new tab)
s2 · cryptopolitan.com · News report · English · Inspected · Shares an underlying account with another listed source
How the sources were read, and where the events happened
English column by Julia Angwin for The New York Times, republished by the Oman Observer on 25 September 2023; read in full on 2026-10-24 by the research agent (an AI); no translation required. The columnist quotes the actor directly ('he told me') and relays IBM's statement. Applies to s1.
English article of 1 July 2023 read in full on 2026-10-24 by the research agent (an AI); no translation required. It does not cite its source; it carries the same witness's account and shares his independence group. Applies to s2.
Event countries: Unknown. Affected-person countries: United Kingdom. Court countries: Unknown.
The clone was sold to websites whose locations are not given, so no event country is recorded. Both sources describe him as a British voice actor, recorded as the affected person's country.
Reviewed for publication 2026-10-24: Published as an ongoing adverse experience: a named British voice actor says an AI clone of his voice, built from an old IBM recording and sold to websites, was used without his consent and cost him jobs. The AI system's generated speech in his voice is the reported cause; the account is his, IBM's response is included, and the outcome is unknown.
People described
A British voice actor, Greg Marston, who gave named interviews about the case
People reported harmed in this case
1 person
0 AI participants · 1 other person harmed
One person: the voice actor whose voice the clone reproduces, by his account. He did not use the AI system.
Counted once within this case. The same person may appear in other cases. This count does not establish AI causation.
Cite this case
Compiled per our published methodology: verification statuses, sourcing standards, and corrections process.
APA
NOPE. (2026). British voice actor says an AI clone of his voice, built from an old IBM recording and sold to websites, has cost him jobs. AI incidents. https://nope.net/incidents/2023-british-voice-actor-says-ai-clone-of-his-voice-made-from-old-ibm-recording-was-sold-to-websites-and-cost-him-jobs
BibTeX
@misc{2023_british_voice_actor_says_ai_clone_of_his_voice_made_from_old_ibm_recording_was_sold_to_websites_and_cost_him_jobs,
title = {British voice actor says an AI clone of his voice, built from an old IBM recording and sold to websites, has cost him jobs},
author = {NOPE},
year = {2026},
howpublished = {AI incidents},
url = {https://nope.net/incidents/2023-british-voice-actor-says-ai-clone-of-his-voice-made-from-old-ibm-recording-was-sold-to-websites-and-cost-him-jobs}
} Related cases
Irish voiceover artist says an AI text-to-speech voice made from her recordings was resold online and used in ads and products she never booked
Remie Michelle Clarke, a voiceover artist in Dublin, told Bloomberg Law in April 2023 and Marketplace in June 2023 that a synthetic voice built from recordings she had made for Microsoft's text-to-speech software was being sold by a third-party AI voice company, Revoicer, as a voice called Olivia that customers could make say anything. She said she learned of it when a fellow actor asked whether another voice job had paid her well, that she had since found a number of adverts and a couple of tech products using her voice, and that buyers choosing the AI voice was affecting her smaller jobs. She said she was completely shocked. Microsoft told Marketplace that it informs voice talent and obtains their legal consent to create synthetic voices; it did not address the resale. No response from Revoicer and no legal action are reported in the inspected sources.
Incheon high-school student jailed for AI deepfake nude composites of his teachers posted online; a civil court later orders him to pay one teacher damages
In July 2024 a high-school student in Incheon, South Korea, composited the faces of two female teachers at his school and of others, including an academy senior and an instructor, onto nude photos and posted them on social media, according to the charges as reported by Yonhap, the Hankyoreh, the JoongAng Ilbo and the Kyunghyang Shinmun, which describe the technique as AI-based deepfake technology. He had also secretly photographed teachers in class. He was expelled after a teachers' rights committee hearing. On 27 August 2025 the Incheon District Court gave him an indeterminate prison term of one year to one year six months and took him into custody; the court said that as a result the victimised teacher had left teaching and had not returned. On 19 December 2025, after he had turned adult, the appeal division raised the sentence to three years, calling the offence close to 'character assassination' and the teachers' mental shock considerable; OBS reports the sentence is final. On 21 October 2026, according to the Incheon Teachers' Union, a civil judge at the Incheon District Court ordered him, now 20, to pay one teacher 17.7 million won of the roughly 72 million won she claimed.
Jeju: police refer a middle-school student over secretly filming teachers and making sexual deepfake composites; teachers dispute victim count and legal aid
Jeju Dongbu police announced on 23 September 2026 that a student at a middle school on Jeju had been referred to prosecutors without detention on 17 September for violating the Sexual Violence Punishment Act (filming with a camera or similar device). The student's phone was voluntarily submitted by his guardians and forensically examined; the examination found footage relating to some teachers and some sexual composites, and the student admitted the conduct. News1 reported that the student was booked for filming a teacher's body with his phone during a school field trip around mid-July 2026, that the offending had continued for several months with multiple victims, and that police found no indication the composites had been distributed. Police told Seoul Shinmun there were more victim teachers than the one first reported but withheld the number to avoid secondary harm, and that no student at the school was a victim; the Jeju teachers' union questioned whether the full scope of harm had been established. According to the union, two months on the victim teachers were receiving hospital treatment and counselling while teaching daily, gave police statements on 14 September without a lawyer, attended the regional teachers' rights protection committee on 17 September without institutional support, wrote the criminal complaint themselves, and were offered 3.3 million won in legal fees for all of them together, payable after the case ends; two requests to meet the superintendent were not accepted. On 29 September one of the victim teachers gave a press interview at the Jeju Teachers' Union office. She said that early in the investigation she had seen on the student's phone material showing other teachers and current students, had made a list of the people and types of harm, and that the material shown to her at the police victim examination was considerably less than what she had seen; Yonhap-chain outlets quote her as saying the students she recognised numbered 10 to 20 and that a graduate had suffered worse deepfakes. She said investigators asked the school only to identify the perpetrator and never requested material to identify the students in the footage, and that the school sent parents a letter premised on there being no student victims. Police replied that phone forensics had been thorough, that every act beyond social norms had been identified and proven, and that one student victim had been confirmed, a student at a different school. The teachers were notified on 22 September that the case had gone to the prosecution and on 24 September that it had been sent to the court's juvenile division as a juvenile protection case, before they had submitted further evidence and a victim opinion. The teacher said the original complaint had also asked police to examine whether physical contact in 2025 was forcible molestation and whether repeated approaching and waiting in 2026 was stalking, and Halla Ilbo reports that such conduct is said to date from 2025. Yonhap-chain reports describe the charged victims as five teachers. The same day the education office briefed on legal-aid rules (up to 6.6 million won per instance when a teacher is sued, 3.3 million won per case when a teacher files a complaint; office lawyers cannot act as a teacher's private counsel), said it had linked the teachers to the national Digital Sexual Crime Victim Support Center, announced an integrated support system through a hotline, and its section head said the office was sorry that early support had not been felt. The KTU Jeju branch (28 September) and the Jeju Teachers' Union (14 and 29 September) demand a victim-support system, per-teacher legal fees paid in advance and a dedicated support team. On 6 October 2026 Seogwipo Broadcasting reported, citing people in education and a tipster, that the education authority had finally rejected the victim teachers' application for lawyer fees for an additional criminal complaint, saying in a written opinion that with a police investigation already under way a further complaint could hardly stop or prevent the infringement of their teaching, so the support requirements were not met; critics contrast this with full legal-cost support in earlier sexual-offence cases.
Thanh Hóa, Vietnam: police say nearly 20 people were threatened with fabricated sexual images of themselves unless they paid; one demand was 1.5 billion dong
Thanh Hóa provincial police said on 3 October 2026 that their cybersecurity and high-tech crime division had, since the start of October, received reports from nearly 20 people threatened with extortion using sexual images and videos spliced from their own photos, which senders threatened to publish unless money was transferred. One victim was asked for as much as 1.5 billion dong, and officials, including commune-level leaders, were among the targets. The police release says the material was made with technology; Báo Thanh Hóa, reporting information compiled from the same police division, says the senders used AI and deepfake technology to put victims' faces onto sexual images. No payment, arrest or named victim is reported.
If you or someone you know is struggling, free and confidential support is available. Find a helpline near you at Signpost.