Artificial intelligence is evolving so fast that it feels like it was launched a long time ago. But it’s only been a year and a half since we met ChatGPT. Back then, it was blind and mute. Today ChatGPT already has a voice, and also listens to what we say. One more step towards the humanization of AI.
A speaker-shaped button in the ChatGPT app allows us to give commands to OpenAI’s artificial intelligence with our voice. It will also respond to us with its own voice, preselected from among four different actors.
OpenAI ran into trouble a few months ago when it presented a demo in which the voice sounded suspiciously like Scarlett Johansson, without her consent. The actress was very upset, so it seems that OpenAI has now chosen four actors who participated voluntarily, and have been paid for their work.
This is how voice works in ChatGPT
At the moment, this new chatbot feature is only available to a group of users with the ChatGPT Plus subscription. It will be available to all paying users in September.
As explained by OpenAI in Xthey have trained Voice Mode with over 100 members of their tester teams from 29 countries, in 45 languages. So it’s not just going to be available in English.
They have also added guardrails to prevent the default four voices from being changed.deepfakes being carried out, and voices being used for illegal or unethical acts.
The AI behind the voice mode in GPT-4o. According to OpenAI, it can speak natural language, and stop when interrupted. It also is able to detect and respond with humor and sarcasm.
However, it still can’t do things that OpenAI showed in the demo, such as detecting photos and videos. This is probably for security reasons.
ChatGPT now has a voicealthough for the moment, only for a privileged few. AI is becoming more and more human, and it will soon become indistinguishable from us…
Known how we work in ComputerHoy.
Tags: Artificial intelligence
Add Comment