Can AI clone a voice convincingly from a 30-second sample ?
Cast your vote — then read what our editor and the AI models found.
ElevenLabs put broadcast-quality voice cloning on a SaaS dashboard. Audiobooks, dubbing, and scam-call detection all changed shape.
Current AI systems are capable of cloning a voice convincingly from a relatively short audio sample, including a 30-second clip. This is achieved through the use of deep learning algorithms, such as waveform models and neural vocoders, which can learn the patterns and characteristics of a voice from a limited amount of data. These models can then generate new audio that mimics the original voice, including its tone, pitch, and rhythm. The quality of the cloned voice can be remarkably high, making it difficult to distinguish from the real thing.
— Enriched May 9, 2026 · Source: IEEE Spectrum — https://spectrum.ieee.org/ai-clones-human-voices
Gallery
No images yet — upload one below to start the gallery.
Disagree? Post your comment below.