OpenAI previews synthetic voice creator, Voice Engine

The exemplary tin create life-like audio from a 15-second clip and a matter prompt.

OpenAI launched sound capabilities successful ChatGPT past September. Now, nan institution is previewing a exemplary called Voice Engine, which tin usage a azygous 15-second audio clip and matter punctual to make longer audio. OpenAI boasts that Voice Engine produces life-like voices pinch inflection and tone, alternatively than a robotic drone.

According to OpenAI's blog post, Voice Engine was first developed successful 2022 to powerfulness ChatGPT's Read Aloud feature arsenic good arsenic text-to-speech. Since then, OpenAI has tested Voice Engine successful a number of different scenarios: for children and non-readers; for non-verbal group and group who person different mislaid their voice; and translation. For each of these cases, OpenAI has collaborated pinch companies successful these respective spaces.

OpenAI spends a chunk of nan blog station assuring that Voice Engine is built safely — and while it doesn't mention nan infamous robocall of Joe Biden explicitly, it's implied: "We admit that generating reside that resembles people's voices has superior risks, which are particularly apical of mind successful an predetermination year," nan station states. "We are engaging pinch U.S. and world partners from crossed government, media, entertainment, education, civilian nine and beyond to guarantee we are incorporating their feedback arsenic we build."

The Biden robocall was apt made pinch package by ElevenLabs, not an OpenAI product, but Voice Engine mightiness clasp nan aforesaid capabilities. At nan extremity of nan post, OpenAI states that owed to its "approach to AI information and [their] voluntary commitments" (committing to safety, security, and trust), it's not releasing Voice Engine wide correct now.

"We dream this preview of Voice Engine some underscores its imaginable and besides motivates nan request to bolster societal resilience against nan challenges brought by ever much convincing generative models," nan station states. OpenAI past calls for a phase-out of voice-authentication arsenic a information measure; policies to protect nan usage of people's voices successful AI; education; and much exertion to place inauthentic voices — each because of its ain technology.

Read much and perceive to examples of Voice Engine successful OpenAI's blog post, Navigating nan Challenges and Opportunities of Synthetic Voices.

This newsletter whitethorn incorporate advertising, deals, aliases connection links. Subscribing to a newsletter indicates your consent to our Terms of Use and Privacy Policy. You whitethorn unsubscribe from nan newsletters astatine immoderate time.