Modulating custom voices: The commercialization of voice cloning

• Author:
• Author name

  Quantumrun Foresight
• January 20, 2023

Artificial intelligence (AI) is capable of creating synthetic voices that are replicas of the original voice. Not only can AI mimic the intonation but also the emotional cadences over time. This development can benefit the entertainment and healthcare industries. However, this innovation has a dark side.

Modulating custom voices context

Voice cloning technology began with basic speech synthesizers, which can convert text to human speech. For example, Google Translator can read aloud a text in a foreign language after translating it. Text-to-speech voice conversion helps create audio content so that producers never have to reach out to voice actors. 

Overdub is a service used by podcasting and radio industries for its ultra-realistic voice cloning ability. Other companies, like Replica Studios, Lovo, Voice-mod, Resemble Ai, DeepZen, Sonantic, and VoiceID, can do the same thing. These tools enable people to make any celebrity or person with publicly available voice samples "say" anything they want. 

In short, any voice can be converted into another person's voice. The only requirement is an hour-long, high-quality recording of the target's voice, allowing AI to generate its model correctly. Once processed, you can clone unlimited speech to your target voice without sacrificing intonations, cadence, and particular vocal emphasis.

Multiple companies are releasing new algorithms to create a more sophisticated deepfake voice technology, such as China's Baidu Deep Voice Research team. Their technology can clone voices within 30 minutes if provided with sufficient data. Similarly, Adobe launched a program that can mimic voices within 20 minutes. These advancements are being made for positive applications, including giving a voice to patients with hearing impairments, Parkinson's disease, or ALS (Amyotrophic lateral sclerosis). 

Disruptive impact

Soon we might be able to deploy personalized synthetic voices for every consumer. Imagine hearing the comforting voice of a loved one who has passed away or the playful voice from your favorite childhood cartoon. However, in this era of misinformation, synthetic voices can also be used to influence or mislead consumers.

The voice is like a fingerprint—each person's is different. As such, some governments and financial institutions use voice authentication to access private assets. In addition, people usually rely on their natural ability to distinguish the voices of friends and family when they cannot see them. These circumstances allow voice scammers to access people's personal information or money.

Because of this threat, several countries' law enforcement agencies created rules for artificially synthesized voices. In 2019, the US passed DEEP FAKES (The Defending Each and Every Person from False Appearances by Keeping Exploitation Subject to Accountability Act). This law penalizes anyone who uses any kind of deepfake technology (video, audio, images) for fraud. 

In the European Union (EU), the Code of Practice on Disinformation was updated in 2021 to urge big tech firms like Alphabet, Facebook, and Twitter to take stricter measures to combat any synthetic media used to spread disinformation. Failure to comply can result in fines of up to 6 percent of the company's global revenues. In 2022, the Cyberspace Administration of China released proposed guidelines to manage the production and spread of fakes created using deep learning or virtual reality technologies, such as artificially altering face or voice biometric data.

Implications of modulating custom voices

Wider implications of modulating custom voices may include: 

• Actors or celebrities renting out their voices to companies, eliminating the need to do actual studio recordings. 


• Voice actors being replaced by synthetic voices that can replicate any intonation or modulation. 


• More companies advancing synthetic voice technologies to support the entertainment and gaming industries.


• People using synthetic voice sites to record the voices of deceased loved ones and keeping them as keepsakes.


• Increased incidents of voice scammers, resulting in more identity and data theft and financial crimes. When combined with AI automation and Natural Language Processing systems, it may be possible to scale up phone scams against a much larger population (likely by impersonating family members).

Questions to comment on

• If you've tried synthetic voice software, how accurate was it?


• How else do you think this technology is going to be used?