Voice clones of non-native English speakers rated more human: what it means for companies
A NeurIPS 2026 paper cloned the voices of 86 non-native English speakers using three voice-synthesis systems. In 4,000 human ratings, the clones were judged warmer, more authoritative and more human, which raises trust risks for voice channels.
Source: Clones de voz de falantes não nativos de inglês avaliados como mais humanos que os… (x.com). Text prepared with AI from this source.
What happened and what to do
The NeurIPS 2026 paper cloned the voices of 86 non-native English speakers using three TTS voice-cloning systems. In 4,000 human ratings, the clones were judged warmer, more authoritative and more human. The study offers evidence on how listeners perceive current systems, including the effect on trust.
In practice, a company that receives calls, voice messages or approval requests by audio cannot treat how natural a voice sounds as proof of identity. One path is to review processes that rely on voice for authentication, require out-of-band verification for payments and account changes, and monitor voice channels for suspicious patterns. It is also possible to build a private test bench that evaluates synthetic-voice detection on the company's own recordings, respecting consent and control over its voice data.
How the consultancy can help
Diagnosis of exposure to synthetic-voice fraud across the company's channels, such as customer service, collections and internal approvals. Then design and deployment of out-of-band checks, voice-channel monitoring and consent policies for voice data, operated on the client's infrastructure.
Next step
To assess your company's case, send a short description of your voice channels and current verification processes. We will reply with a scoped proposal for diagnosis and implementation.
Consulting for your project
Infrastructure review, deployment and ongoing operations, with scope and pricing defined in the proposal.
Quoted per project
Request a proposal