Deepdub CEO on the Power of AI Voice Clones to Reshape Streaming Content

Generative artificial intelligence (AI) is transforming nearly every industry it touches.

And when it comes to sectors like entertainment and media, where the production of content is the name of the game, the impact of generative AI’s capabilities can’t be overstated.

“Generative AI delivers a very simple way to generate an end product,” Ofir Krakowski, CEO and co-founder at Deepdub, told PYMNTS during a conversation for the “AI Effect” series.

Krakowski, a native Hebrew speaker from Israel, highlighted the fact that a significant portion of audiovisual content on the internet is in English. This creates a language barrier for non-English speakers but is also an opportunity for companies like Deepdub, which utilizes generative AI tools, to bridge that gap.

“Sixty percent of the audiovisual content out there is in English, and around the world a lot of people don’t understand English. This means that 60% of the content globally is not accessible to most of the people around the world … what we are doing [with Deepdub] is democratizing the content,” he said. 

Deepdub specializes in localizing entertainment content and generating multilingual voice clones using AI, showcasing the practical applications of generative AI’s advanced applications, which include having voice actors and celebrities narrate a movie in various languages without losing the nuances and tonal inflections that make their voices special and recognizable.

Still, as Krakowski noted, there is an ongoing need to balance technological advancement with ethical considerations to avoid exploitation. For instance, within the voice-acting community, there are long-standing concerns around the usage of their recordings in projects without proper consent or compensation, tensions which were brought to a head this summer with the Hollywood strikes.

Fully Automating Legacy Processes

AI’s applications within the entertainment industry are fundamentally changing the ways in which content is created and distributed. 

“The reason 60% of global content is not localized right now is because it’s not economically viable doing it with humans — there aren’t as many voice actors around the world as we need,” Krakowski said, adding that localization takes multiple steps beyond just direct translations, including the contextualization of idioms.

“It’s a multimodal problem that we are trying to solve the full workflow of in an automated and simple way,” he said. 

That’s why Deepdub recently launched a Voice Artist Royalty Program. By enabling voice talents to monetize their likeness through AI-generated content, the royalty program ensures fair compensation, fostering a symbiotic relationship between technology and human creativity. Krakowski said that the program aligns with the company’s commitment to ethical and legal practices, while also expanding its goal of democratizing formerly English-only content.

As AI-generated content gains prominence, legal and ethical concerns emerge.

Krakowski emphasized the importance of ensuring that AI tools are built on ethical and legally acquired data. This, he argued, safeguards companies from potential legal challenges related to intellectual property.

“When it comes to the companies that are using AI, not the companies that are building it, what they actually care about is ensuring that they won’t get sued and that if they create something using generative AI, the end result has their intellectual property (IP). The tools they use need to be enterprise-grade,” Krakowski said.

Creative Coexistence

Despite AI’s capabilities, Krakowski noted that certain aspects of creativity, such as translating idioms and creating jokes, remain challenging for AI. He expressed optimism about the continued relevance of human creativity in media and entertainment, particularly in areas where AI falls short.

Where AI can play a compelling and attractive role, Krakowski said, is in helping entertainment businesses to “monetize content that is otherwise not economically viable to monetize” by effortlessly localizing programming.

“I believe we’re going to see TV channels generated wholly with AI, topic-based news channels where there is no human in the process,” he added, emphasizing that it is still a little way away, and that a fully AI-generated scripted show or movie is likely not possible, at least without human involvement.

“We can see front-facing avatars, but you cannot generate a drama series with a lot of motion, complex motion for a long format, this will take some time,” Krakowski said. “Imagination and new creation is not something AI is capable of … Right now, it is just mimicking what it has learned.”

As for the future of entertainment within a generative AI world, while challenges and regulatory considerations persist, the evolving landscape promises a future where AI and human creativity can coexist, reshaping the way we consume and produce entertainment content.