What The Tech?: My Own Voice And Chat GPT
CHARLOTTE – Chat GPT, or chat.openai.com is now the world’s fastest-growing website. A new report shows that in March, the website attracted some 1.6 billion visits and the number grows each month. If you’re unfamiliar with how it works, you simply type a question or command, and Chat GPT can answer questions and create content just like humans. What would happen if you combine Chat GPT with audio-replicated software? Could it not only read but sound like a human is saying what Chat GPT comes up with?
I recently looked at a voice recording platform from The Acapela Group that stores voices in a “voice bank” for use later. You record yourself saying random phrases and words. After processing, someone can type out what they want to say, and have it read back in their voice. Its primary function is to help patients who’ve lost the ability to speak. ALS patients for example, can type out what they want to say, and rather than a computerized voice reciting it from a phone or device, the voice will be their own.
For instance, someone diagnosed with ALS can record themselves saying 50 or more phrases into a microphone. Once the recording is complete and processed they can type out what they want to say and have it read aloud in their voice over a computer, device, phone, or tablet.
Helpful and reassuring for patients who lose their voice.
What if you could use the technology for other reasons? Maybe create a presentation using Chat GPT, and have My Own Voice, present it as an audio recording or podcast? I tried this asking Chat GPT to create an explanation and introduction to What the Tech? using the information at www.whatthetech.tv. In a matter of seconds, Chat GPT had created a full presentation for a room filled with people. I then copied the Chat GPT text into My Own Voice. The audio playback was unnatural for a presentation and sounded like a computer, albeit in my own voice.
This isn’t likely to happen with My Own Voice, but other companies offer artificial voice technology to create deep fakes. There are dozens of deep fake apps in both app stores (though I don’t recommend downloading them). Apple is expected to release its own version in an upcoming update where Siri can read anything you type in your own voice. Will it sound natural? Could someone confuse an AI voice with a human talking on the phone? The technology already exists. The question is how soon will it come to your phone, tablet, or computer?