The capability of artificial intelligence does not end in simulating a human's facial expression. Over the past years, other companies have focused on bringing unique AI-generated voices that exactly mimic a human speech.
For instance, AI voice firm Sonantic recently unveiled its interest in audio deepfakes. In this regard, the startup developed a machine-operated voice that could tease or flirt with someone.
Sonantic Creates a Flirtatious AI
According to a report by The Verge, Sonantic managed to come up with a unique synthetic voice that could communicate with humans. By incorporating non-speech sounds, the expressions become more authentic as they imitate human-like behaviors.
In an interview with the tech site, John Flynn, the co-founder of Sonantic, said that they have ended up choosing love as the "general theme."
The company's CTO added that the goal for the company is bringing subtle emotions by means of modeling them.
According to CEO Zeena Qureshi, the company relies on the "Photoshop for voice" software, which helps them tweak the AI's mood when delivering the phrases or sentences.
The software is also helpful in selecting an ideal type of AI voice that was originally taken from real-life actors. Sonantic said that its software offers more "in-depth" customization than its competitors use.
Interestingly, the AI voice could also deliver emotions including joy, sadness, fear, anger, and more, not to mention its capability to tease and flirt with humans.
"I think that's the main difference - our ability to direct and control and edit and sculpt a performance," Flynn said in Sonantic blog.
He added that they have partnered with entertainment studios mostly under triple-A. He mentioned that they also established a collaboration with Mercedes for its digital assistant for the cars.
There's also a "director mode," which gives more detailed adjustments in pitch and intensity of delivery. Other than that, it also adjusts laughs, growls, and even breaths.
Audio Deepfakes Are Stellar Creations
For Sonantic, this breakthrough was an achievement in creating fresh audio deepfakes through machine learning. To get the best out of the AI voices, further polishing is still needed, specifically done manually.
After all, synthesizing AI voice can bring advantages to several industries, including gaming and entertainment. It makes sense why developers alter the type of voices. This attention-grabbing process makes voice acting unique and suitable depending on the situation.
While deepfake voices can be used for imitating human speech, they can also be used with bad intentions.
Deepfake Vocals
According to Tech Times, the hackers could deceive the bank employees into knowing that they were associated with the bank. There was an ongoing business transaction at that time.
The bank robbers reportedly used AI-enhanced voice simulation to stage a $35 million heist.
Regarding deepfakes, Chinese regulators intensified the crackdown by proposing new restrictions for the use of AI technology. The authorities also aimed to promote socialist values on several platforms.
This article is owned by Tech Times
Written by Joseph Henry
