As the development of generative artificial intelligence (AI) has made it possible to synthesize even voices, a ‘voice watermark’ technology has emerged to identify them.
Tech Crunch, a professional media outlet, reported on the 1st (local time) that Resemble AI’s generated voice watermarking technology, ‘PerTH’, was proposed as a way to resolve voice fraud concerns.
AI-generated voice technology is raising concerns that it can be misunderstood as if it were actually spoken by politicians or celebrities, as well as being abused for voice phishing or fraud.
Watermarking is the technology of imprinting an identifiable pattern to identify the source of an image or video. Conspicuous watermarks such as logos and images are common, but there are also many watermarks that are invisible to humans. It hides the pattern at the pixel-by-pixel level so that the human eye looks unmodified, but the computer can see it.
Similarly, in audio, there is a way to encode sounds that are inaudible to the human ear. However, this inaudible watermark encoding method has a disadvantage that the voice watermark may disappear even if the media is slightly modified.
For this reason, Resemble AI solved this problem by using PerTH.
PerTH is a machine learning module that analyzes the sound waves of the voice and automatically discriminates between the sound waves that are smaller than the sound waves.
Watermarks are not only difficult to distinguish from actual sound, but also difficult to remove because they are closely linked to voice information of a similar band. On the other hand, watermarks can also be easily identified with PerTH.
Resemble AI is also an AI startup that provides generated voice services that replace human voices in dubbing and audio book services. Therefore, they said they came up with such a device to prevent their services from being misused.
Resemble AI announced that it will soon provide PerTH watermark service to all customers.
Reporter Juyoung Lee juyoung09@aitimes.com
