Technology.- OpenAI, a leader in artificial intelligence (AI), announced its new Voice Engine model this Monday, which is capable of replicating human voices from just a 15-second audio sample in an “emotional and realistic” way. Despite this technological achievement, the company has decided not to release the technology to the general public, citing concerns about the risks associated with it.
“We recognize that creating speech that mimics human voices poses serious risks, especially in an election year,” ChatGPT's creators said in an official statement.
Voice Engine represents a significant advance in the field of AI by converting text into speech to create synthetic voices. According to OpenAI, the technology will help people who have difficulty with verbal communication, such as those who need to use their voice to make phone calls, and allow content creators to translate their work into other languages without losing their own accents.
An example provided by the company shows that an Anglo-Saxon's voice can be translated into Spanish, Mandarin, German, French and Japanese while preserving the person's native accent.
Currently, only a few companies have access to this technology, including Age of Learning, Heigen, Dimaki, Livox and Lifespan.
“Given the potential for artificial voice misuse, we are taking a cautious and informed approach ahead of wider distribution,” the company explained in its statement, highlighting that cooperating companies have agreed to its usage policies. These policies include prohibiting the use of one's voice without permission and informing users when they interact with artificial voices.
OpenAI has implemented additional security measures, such as watermarking, to identify the source of any audio generated by the Voice Engine, as well as active monitoring of its usage.
OpenAI's announcement comes at a time when cases of robocalls featuring cloned voices of prominent politicians such as President Joe Biden have already been reported in the context of the current election year in the US.
In addition to ethical and privacy implications, another major problem associated with AI-generated synthetic voices is their potential use by fraudsters to access bank accounts using voice recognition.
This may interest you:
The microwave glitch
