By now, we’ve all heard about the potential that artificial intelligence has to improve the world around us. Just look at the traction that chatbot programmes such as ChatGPT are generating worldwide.
Used correctly, A.I. could double business productivity, employee efficiency and help enhance and automate workflows globally. Its potential and scope for innovation is truly limitless… and a lot of different businesses have started taking notice. The world has started to experiment with how the tech could be applied and improve day-to-day life for people. However, like with any new tech, there’s always a flip side to the incredible innovations they preach.
The risks and potential negative applications A.I. have when used in the wrong hands could be troubling to say the least. In this article, we’ll look at the double-edged sword that is artificial intelligence and the potential benefits and risks that emerge from utilising it within the Podcasting industry.
From assisting in writing engaging narrative scripts for listeners to follow to actually helping within the editing suite, A.I.’s potential applications within the podcasting industry are truly limitless.
One thing that is getting better seemingly with every passing week that we could soon see utilised within podcasting is Digital Voice Cloning.
What is it, you ask? Voice cloning is a method of recreating a person’s voice digitally.
How does it do this? The program, using A.I.-backed software, can generate a voice from a databank of recordings which it can then use to replicate the voice with all of its unique characteristics – including pronunciations and accents. The bigger the database the A.I. has to use (i.e. the more voice recordings it has to scan over), the better constructed and more accurate the voice will be.
How could this be applied? This technology is generating increasing popularity amongst podcasters and editors with its potential applications and problem solving solutions. Say that, for instance, you have a prestigious and time-poor businessman as a guest on your podcast and something important they said during a recording was inaccurate. Rather than bothering that guest who is busy running their business with the hassle of a re-record, with a voice clone you have the potential to correct the mistake without needing to cut any material.
Voice cloning also has a handful of other applications that could revolutionise not only the podcast space, but also the marketing industry. With the ability to generate a voice clone anywhere in the world via a handful of one-off recording sessions, the future could see celebrities, voice actors and similar generate their own voice clones which they could then market and sell as an asset to brands and businesses.
How could this be abused? Whilst A.I.’s capacity for innovation are exciting, the potential downsides to a technology with this power are worrying. For instance, with the addition of the digital element comes the increased likelihood for synthetic media and deepfakes.
What is a deepfake? A deepfake is where a person’s likeliness (in this case their voice) is used to create fake and manipulative content with the express intention to deceive.
What could this lead to? Whilst this could understandably defame and damage the reputation of the real person behind the voice clone, the larger implication tech like this could have on the mass spread of fake news and paranoia is the true distresser here.
Think about it… when you hear the voice of a person you trust tell you something, you’re inclined to inherently believe them – right? Imagine if the voices of trusted individuals such as news readers, government officials and celebrities got into the wrong hands.
Key Takeaways: The tech behind voice cloning, as it stands, isn’t good enough yet. Most listeners would be able to tell that a voice has been altered or re-created as the emotion, pace or the tone and inclination of the person’s speech is slightly unusual or off.
However, with this technology getting increasingly smarter and harder to notice – the gap between being to tell whether the voice is digital or real is closing, leading to the increased risk of misuse.
Ultimately, it will be up to the end user and how they use the A.I. driven technology that will affect the impact it has on the actual content. A.I., like any other technology, is neutral and morality is a grey construct. The only thing that opens it up to malicious use is people abusing it to do things it shouldn’t.
Hence, the conversation around this emerging tech should shift towards procedure, practice and law in place around the things we can control. If people are going to abuse the technology and create fake content with the express content to mislead, then how do the platform distributors, such as Spotify and Apple, detect and filter this content and deconstruct their message? How do we establish that voice clones are the property of the creators themselves, and that they are simply lending or renting their voice out to people? Will this tie into the wider conversation around NFT’s and the ownership of digital content?
Who knows? The only thing that for certain is that the technology isn’t going anywhere… It might not be fully up to standard yet, but with A.I. constantly improving and learning – a future where digital and real voices are intertwined might be closer than we think.
