The Human Voice in the Age of AI

• Courtesy of Pocholo De Leon Gonzales, founder and CEO of CreatiVoices Productions.

Artificial intelligence no longer just generates text, images and video. It can now laugh, whisper, pause, change emotions and reproduce a person’s voice from only a few seconds of recorded audio.

The arrival of ElevenLabs V4 shows just how quickly voice AI is advancing. But for Pocholo De Leon Gonzales, founder and CEO of CreatiVoices Productions and known as the AI VoiceMaster of the Philippines, the bigger story is no longer whether machines can sound human. It is how the technology can be used meaningfully and ethically.

ElevenLabs introduced V4 as its latest generation of speech technology, designed to produce more expressive and context-aware speech. The technology can respond to instructions involving emotion, pacing and delivery while supporting more than 90 languages.

Its voice-cloning capability also shows how low the technical barrier has become. ElevenLabs says Instant Voice Cloning can work with as little as 10 seconds of audio, allowing a short voice sample to generate a synthetic version capable of producing entirely new speech.

For Gonzales, who has spent three decades working in voice acting, broadcasting, training and production, that capability represents enormous potential but also clear ethical risks.

“I am not against voice cloning. I am against unauthorized voice cloning,” Gonzales said.

The distinction becomes more important as artificial voices become increasingly difficult to distinguish from recordings of real people.

Voice recordings are everywhere. Interviews, podcasts, livestreams, social media videos and short online clips can provide material that increasingly sophisticated AI systems can use to reproduce elements of a person’s voice.

The same technology that can help creators produce content faster can also be abused for impersonation, fraud and misinformation.

Consent, Gonzales said, must therefore remain central to voice AI. A person’s voice should not simply become another piece of digital content that can be copied because it is publicly available.

Permission, attribution, licensing and compensation will become increasingly important as voice cloning becomes part of mainstream content production.

“Technology itself is not the enemy. Abuse is,” Gonzales said.

But focusing only on the risks would also miss what the technology can do.

Used productively, voice AI can significantly lower the cost of producing educational materials, audiobooks, training programs, customer-service systems, games, animation and localized content.

A Filipino teacher could create lessons in several languages without recording every version individually. Businesses could produce customer information for different markets, while content creators could make material accessible to audiences who speak different languages.

Voice AI could also improve accessibility and create new ways for people to consume information through conversational assistants and audio-based services.

For the Philippines, the possibilities go even further.

The country has numerous languages, regional accents and distinct ways of communicating. Gonzales believes Filipinos need to participate in developing the datasets, pronunciation models and cultural context behind these technologies if AI systems are expected to represent Filipino voices properly.

Gonzales has taken that approach himself, experimenting with AI voice technologies, cloning his own voice and contributing Filipino speech, pronunciation and cultural context to emerging synthetic voice systems.

The shift could also change what it means to become a professional voice artist.

Voice actors may eventually work not only as performers but also as licensed voice providers, AI voice directors, trainers and specialists who help machines understand timing, emotion, delivery and cultural context.

Voice direction could evolve as well. Instead of directing only human performers, future directors may also guide AI systems using prompts, emotional instructions, timing and performance cues.

The profession will change, but that does not necessarily mean the disappearance of the profession.

Gonzales has consistently argued for a different way of looking at the disruption: not AI versus voice actors, but AI working with voice actors.

That approach also shapes his advocacy around productive AI use.

For educators, creators, entrepreneurs and businesses, the opportunity is not simply to automate existing work but to use AI to reach wider audiences, develop new services and create forms of content that were previously difficult or expensive to produce.

Gonzales has expanded that advocacy through initiatives focused on AI education, entrepreneurship and literacy, with the goal of helping Filipinos become more capable users and creators of artificial intelligence.

The conversation, then, has to move beyond the familiar question of whether AI will replace people.

A more useful question is how industries can adapt as tools such as ElevenLabs V4 make synthetic speech faster, more expressive and easier to produce.

For the Philippine voice industry, the answer may lie not in resisting the technology, but in learning how to use it without surrendering consent, creative ownership and the value of the voice behind the machine.