Pages

This Artificial Intelligence Platform Can Create Deepfakes Of Your Voice For Commercial Use

Veritone has recently launched Marvel.AI, a platform where influencers and individuals can create deepfakes of their own voices to license.

It's impressive, if not simultaneously daunting, how far A.I. and deepfake technology has come. The ability for AI to swiftly recognize facial patterns from a series of different photos and seamlessly applying them to those of others is a rather common example, to start off with. The likes of VFX houses such as CorridorDigital are rather well-known in the online space for their work on deepfakes, even rendering one of Tupac for a Snopp Dogg music video. The implications of work such as this in terms of potential security threats and identifying legitimacy are too numerous to discuss in this article. What we can choose to do instead is soldier on and discuss the Marvel.AI and it's inner workings.

Marvel.AI operates much like any other deepfake A.I. would, using a group of samples from a person's voice in order to generate a general baseline for it. Not only will that baseline then produce a monotonous example of said person's evocation, it promises to encompass other emotions and ranges accordingly. Proudly hailing voice to be the next digital revolution, the Veritone website considers options such as Marvel.AI a key component of navigating online content generation moving forwards.

Technology such as this opens up gates for so much more that can be done by the likes of voice actors, celebrities, and even individuals not well versed in the art. Celebrities and influencers without access to proper sound booths or the time to consistently record audio messages for sponsors no longer need to worry. They can easily use the Marvel.AI's generated voice, which will cater to literally whatever dialogue is required. And "literally whatever dialogue is required" is an appropriate question to ask. A lot of online content is simply audio related nowadays. Think about podcasts, audiobooks, lectures, and so on. That, too, is without counting the number of advertisements and endorsements that come with all such media.

Naturally, safety is a major concern in everyone's mind. How do we know that our online voice will stay safe and not be used for nefarious purposes? Well, since the voice library isn't public access, no one can access your samples but you. Along with that, features such as inaudible watermarks, and the ability to claim and protect your voice's usage makes it a good platform to start audio work with.


Read next: A recent survey conducted tells how many people believe in the 5G conspiracies that have made their way all over the place

No comments:

Post a Comment