Mar 29, 06:10 PM

OpenAI's Voice Engine Clones Voices in 15 Sec with 0 Technical Skill, Raises Security Concerns

OpenAI has introduced a new model named Voice Engine, which can generate emotive and realistic voices using just a 15-second audio sample. This technology has sparked discussions about its potential implications and the readiness of society to handle such advancements. Early evaluations of Voice Engine show that it can closely mimic the source speaker's voice from text input and a brief audio sample. However, concerns have been raised regarding the ease with which individuals, with 15 minutes and 0 technical skill, can bypass security measures to clone voices of public figures, suggesting a need for OpenAI to address deployment priorities and ethical considerations.

#OpenAI #Voice Engine

Written with ChatGPT (GPT-4).

Sources

Decrypt@DecryptMedia
3 years ago
New ‘Voice Engine’ from OpenAI Needs Only 15 Seconds to Clone Speech ► https://t.co/w0fuSiPAUU https://t.co/w0fuSiPAUU
Igor Kurganov@IgorKurganov
3 years ago
It took me 15 minutes and 0 technical skill to get past HeyGen's (= @OpenAI's Voice Engine) security to create a voice clone of a public person. It now says whatever I want. OpenAI should just be frank about their deployment priorities rather than apparently duplicitous. Yes,… https://t.co/uppmDU7UsG https://t.co/yIBSlnfo7L
shyamal@shyamalanadkat
3 years ago
early findings from an initial evaluation of Voice Engine, a model that generates speech closely resembling the source speaker's voice from text input and a 15-second audio sample. https://t.co/cvrMrSwJPb

Additional media

Image #1 for story openai-s-voice-engine-clones-voices-15-sec-0-technical-skill

Image #2 for story openai-s-voice-engine-clones-voices-15-sec-0-technical-skill

Image #3 for story openai-s-voice-engine-clones-voices-15-sec-0-technical-skill