Concerns about selling voice: Years of voice training, AI reads in "minutes"

(Dan Tri) - The number of projects is decreasing, competition is increasing and the risk of voice theft is causing many voiceover professionals to face unprecedented pressure from AI.

Concerned about keeping my job and keeping my voice in the face of the AI ​​wave

"I and many others in the profession feel that the number of projects we receive has decreased compared to before," Bich Tuyen (31 years old), who has been working as a voice actor for nearly three years, shared.

That is the reality that Tuyen witnessed recently, when artificial intelligence gradually changed the way audio products were created.

AI voice cloning is a technology that uses machine learning models to create or copy human voices from a certain amount of audio data. This technology is being applied in audiobooks, switchboards, videos, advertising and many digital content products.

With just a paragraph of text and a few operations, users can completely produce a recording without investing in a studio, specialized microphones or a live reader.

According to Bich Tuyen, the pressure on her today comes not only from competition between many good readers but also from a tool that can create products faster and cheaper.

Before facing the pressure from technology, Tuyen experienced the feeling that due to lack of experience, she repeatedly failed to pass the selection round and continuously received shakes of the head from projects.

But then when skills gradually become stronger and work begins to bring in income, the pressure does not ease.

At this time, she faced a more difficult problem than before: how to get customers to still choose the human voice, while AI reads more and more fluently and can create products in just a few minutes?

It was these questions that unintentionally caused Tuyen to fall into a state of uncertainty and disorientation in her favorite job.

Like Tuyen, Thien Trung (30 years old), who has been involved in the voiceover profession for eight years, is witnessing a clear differentiation of the market.

But for him, the worry is not limited to having to compete with a "rival" who can work continuously.

That's the fear when AI has the ability to simulate real human voices. As a result, professionals can lose control of the very sound that is associated with their name, work and identity.

"A voice is a valuable asset. If copied and used, whether for entertainment or commercial purposes, it can affect the career, income and self-esteem of the person being copied," Trung shared.

The investments behind a voice read

In fact, if you just listen to a short advertisement or a smooth audiobook, few people imagine that behind it is a whole process of persistent training.

For Thien Trung, that journey started in 2018 on trendy football fields.

Initially, Trung only saw this as a side job. However, when he met people who were seriously pursuing the profession, he gradually felt that this job could become a long-term direction.

During the first two years, he independently watched every voice training video, signed up for courses and practiced pronunciation, breath control, reading rhythm, pitch or how to handle emotions almost every day.

Only after many classes and thousands of hours of self-practice did the first results appear when he started receiving projects from brands.

If Trung came to the profession from trendy football fields, Bich Tuyen's journey started from his childhood dream.

Having participated in hosting since childhood, Tuyen decided to register for the voice experience experience event when she was looking for a side job in addition to her office job.

At the event, she was selected in the group of three best ad readers and was considered to have a suitable voice for long-term development.

That encouragement became a push for Tuyen to officially return to pursuing her dream.

In November 2023, she received her first project with income. Up to now, after nearly three years, she still maintains her voice training, listens to each recording and corrects the mistakes she often encounters.

However, while professionals take years to build skills and practice every day, AI can create a recording in just a few minutes.

On social networks, it is not difficult to come across individuals and units using AI voices to make audiobooks, tell stories, read news, introduce products or narrate content.

When watching such products, some listeners often comment that the AI ​​voice has a steady rhythm, unnatural accents, creating a "lifeless" feeling and are ready to block products with AI voices.

But on the contrary, many people still accept it because what they care about more is the attractive content and amount of information received.

Therefore, a part of customers also consider AI to be a complete replacement for voice actors because of its convenience, speed and low cost.

Keep the "human quality" and copyright to the voice

Talking to Dan Tri reporter, Mr. Nguyen Phong Anh, an AI expert, said that voice generation tools have developed much more sophisticated than popular imagination.

New generation tools can even simulate small expressions such as talking while smiling, smirking or making swallowing sounds while speaking.

"We often think that AI has no soul. But when listening to a compelling story, viewers can forget that soullessness, even if it is still quite obvious," Mr. Phong Anh shared.

However, AI still cannot fully represent the diversity of voice nuances. Technology has advantages when reading long, relatively monotonous texts, but is limited with recordings that require a flexible, emotional voice.

However, Mr. Phong Anh also believes that professionals should not see this as a permanent advantage.

"It's only a matter of sooner or later for AI to catch up with the nuances of the human voice," he said.

However, in order to retain voice copyright, those working in this profession hope that in the future there will be clearer sanctions to help protect their rights.

According to Mr. Phong Anh, this concern is well-founded because technology makes voice copying easy.

While unauthorized users of voice may face legal issues, those who have been infringed have difficulty gathering evidence and proving damages.

Therefore, experts emphasize that voice-over professionals must proactively protect their rights right from the moment they sign the contract.

Experts recommend that if they agree to license, the voice owner should not sell the rights to use it all at once.

He believes that the parties can agree on the fee for recording clean data, the time and scope of exploitation, the number of minutes of audio created, product profits and renewal fees.