Startup Sonantic has made a breakthrough in the development of sound deepfakes by creating a synthetic voice that is taught to tease and flirt
AI has already learned to flirt, but synthetic speech is still recognized
The company says the key to its promotion is incorporating non-speech sounds, like sighs or derision, into the audio.
: The robot for the first time performed a complex operation without human assistance
The startup has trained its AI models to recreate the half-concealed chuckles and short sighs that give real speech biological veracity.
The quality of voices generated by artificial intelligence has improved rapidly in recent years, but there are still aspects of human speech that which are not amenable to synthetic imitation.
Choosing the theme of love the company tried to create a flirtatious AI, but listeners still notice strange spaces between words and synthetic overtones in pronunciation.
Emotional expressions include anger, fear, sadness, happiness and joy, now also coquetry, shyness, teasing and boasting.
You can adjust the pitch of your voice, increase or decrease the intensity of the presentation, and insert small non-verbal vocalizations such as laughter and breathing.
We previously wrote:
Source: Segodnya
As an entertainment author with over 7 years of experience, I bring a diverse background to my writing. I have a proven track record in the journalism industry. My expertise extends beyond writing, as I have experience producing video content for websites and social media platforms, offering a well-rounded perspective on the entertainment world
