OpenAI‘s Voice Engine, a cutting-edge tool capable of cloning a person’s voice from just 15 seconds of speech, remains in its preview phase more than a year after its announcement. Initially unveiled in late March 2024 as a “small-scale preview” of an artificial intelligence service, the tool was intended to be integrated into OpenAI’s API by March 7, 2024. However, the company has yet to confirm a release date or even if it will officially launch the Voice Engine at all.
The Voice Engine generates speech that closely resembles the original speaker’s voice, providing natural-sounding audio output. It powers voices in both OpenAI’s text-to-speech API and ChatGPT’s Voice Mode. By predicting the most probable sounds a speaker will make based on a given text transcript, the tool can accommodate various voices, accents, and speaking styles. Despite its potential, OpenAI continues to test the Voice Engine with a limited number of “trusted partners.”
OpenAI has been cautious with the deployment of its Voice Engine, focusing on gathering insights from its partners. According to an OpenAI spokesperson, the company is actively learning from these collaborations to enhance the model’s effectiveness and safety.
“We’re learning from how our partners are using the technology so we can improve the model’s usefulness and safety,” – OpenAI spokesperson
Developers wishing to utilize the Voice Engine must secure “explicit consent” from the original speaker and clearly disclose to their audience that the voices are AI-generated. These requirements aim to maintain transparency and ethical use of synthetic voices.
Carlos Pereira, CEO of Livox, speaks highly of the tool’s capabilities, particularly its quality and multilingual features.
“The quality of the voice and the possibility of having the voices speaking in different languages is unique — especially for people with disabilities, our customers,” – Carlos Pereira, CEO of Livox
Safety Measures and Market Positioning
In response to rising concerns about AI-generated content, OpenAI has implemented several safety measures for its Voice Engine. These include watermarking to trace the origin of generated audio, helping to mitigate misuse. The company emphasizes ongoing dialogue about responsibly deploying synthetic voices and societal adaptation to these technologies.
“We hope to start a dialogue on the responsible deployment of synthetic voices and how society can adapt to these new capabilities,” – OpenAI
Voice Engine has been trademarked and priced at $15 per million characters for standard voices and $30 for HD quality voices. This pricing strategy positions it competitively within the growing market for AI voice technologies.
Unfortunately, not all developments have been positive. AI voice cloning emerged as the third fastest-growing scam of 2024, highlighting potential risks associated with this technology. OpenAI remains committed to addressing these challenges through cautious testing and stringent safety protocols.
As OpenAI continues to test Voice Engine with select partners, it remains uncertain whether or when the tool will see a broader release. The company seeks to base its decisions on extensive discussions and small-scale testing results.
“Based on these conversations and the results of these small-scale tests, we will make a more informed decision about whether and how to deploy this technology at scale.” – OpenAI
The delay in launching Voice Engine has not dampened enthusiasm among industry leaders like Carlos Pereira, who praises its user-friendly design.
“It is really the most impressive and easy-to-use [tool to] create voices that I’ve seen” – Carlos Pereira, CEO of Livox
Author’s Opinion
OpenAI’s caution is understandable given the sensitivity and potential misuse of voice cloning technology. While the Voice Engine holds great promise, especially for people with disabilities or in multilingual contexts, OpenAI must tread carefully to avoid ethical pitfalls and ensure the technology doesn’t facilitate fraud or invasion of privacy. Until then, the slow rollout is a necessary step to balance innovation with responsibility.
Featured image credit: FMT
Follow us for more breaking news on DMR