ChatGPT owner OpenAI Inc said it is expanding its venture into text-to-audio AI model Voice Engine on Friday, which can create emotive and realistic voices using text input and a single 15-second audio sample.
Synthetic voice generator was initiated in late 2022, and the current model is in its testing phase. Additionally, Voice Engine is built to replicate or clone an individual’s voice, said the firm.
The technology of cloning an individual’s voice has a higher amount of threats than benefits, given the current global political scenario and poses a huge risk in terms of cyber-bullying.
OpenAI emphasized the possible misuse of Voice Engine and proceeded to deploy this technology with caution and care.
In February, OpenAI launched a new text-to-video AI model called Sora which can create realistic, highly detailed scenes, and imaginative scenes from text instructions.
The AI firm looks to initiate discussions regarding the responsible implementation of synthetic voices and explore ways in which society can adapt to these emerging capabilities.
The firm said, that with this technology, it aims to provide reading assistance to non-readers and children, translate content like video and podcasts, support people with conditions that affect their speech as well as help people to recover their voice.
Earlier this week, Microsoft and OpenAI ventured into a data centre project called Stargate worth about $100 billion, based on a media report that would contain a supercomputer with millions of specialized server chips to power OpenAI’s artificial intelligence.
To read more on Technology, head to the link.
