Part 3: Artificial Intelligence, Voice Generators & Video Creators

This week I met them a lot of wonderful people who complemented my blogs. With some prompted feedback, the general consensus was a specific interest in the AI topics. I mentioned that I did have a few more topics that I wanted to write about and so here is one more.

Last week I watched some videos that programmed celebrity voices to sing covers of the original songs. For example, an original Christina Aguilera song was covered by Britney’s Spears’ voice “Genie In a Bottle”. You can listen to it here. It was quite interesting to hear some of these celebrity covers using other celebrity voices but I did notice some versions had more limitations than others in terms of accuracy. I felt that sometimes the celebrity voices sounded very similar to the original and perhaps it’s just the program that was used, and/or it was the song that was being covered. Further, other limitations included what I called “the cookie cutter effect“ which essentially does not leave room for the singer to put its own spin on the song and so it sounds just like the original. Knowing this, it is quite obvious that these are AI generated covers and not originally performed by the artist themselves as I was already familiar with the songs that I was listening to. But what about prerecorded scripts that are then recorded using AI generated voices?

I watched a video a few weeks ago warning people about the fact that hackers are mimicking people loved ones’ voices to manipulate vulnerable people to sending money. For example, this mother on the video talked about her experience of a hacker calling her and using her young daughter’s voice to ask for money because she is in danger. If she had not known better, she would have sent these strangers money. What’s even more alarming to me is the fact that they had access to her daughter’s voice. My best guess was this was done through social media and so we should really be careful of what we post. You can watch it here. Of course we see some downsides of the development of this technologically, but how can this be beneficial for everyday life?

What Which Ways Can This Be Used In Everyday Life?

It seems to me like the future of the entertainment industry may no longer require celebrities to record their own soundtracks. Even with the AI technology that allows for making commercials, although not very appealing and somewhat scary presently, moving in that direction and refining the technology may help production costs become significantly reduced with making movies, TV shows, etc. With the threat of inflation cutting into profits and budgets, it only seems to make sense that we adapt to this every changing world. Of course, it seems like jobs as we know it may be reduced or extinguished forever, but it doesn’t mean that there will not be other jobs that may be created. Naturally, we fear change because change requires adaptation to our environment and, especially as we get older, uncertainty equates to danger especially if life as we know it changes too quickly to process.

I also predict that celebrities will soon not be required to record their own songs, but rather use a sample of their voices to produce many songs. Just like audio tune, which is a computer program focussed on making celebrities voices sound better upon recording, I think this AI technology will also allow voices to sound better because it will control pitch, breathing, annunciation, etc. This technology will certainly need to be refined in order to make these soundtracks sound like originals and other areas need to be perfected, such as knowing the strengths and limitations of the person whose voice you’re impersonating, how they would possibly perform the song and their little quirks or personalities that shine through the production in order to give it a more human connection to the listener and for greater authenticity.

I always wonder if there will be legal implications of using a persons identity to profit. Of course, the way that videos and voices are advertised will have an impact on protecting the persons identity (e.g. advertising it as Ariana Grande’s impersonation not as the celebrity herself, read more about the apps here.

But how much further should the laws consider to protect the person’a trademark? Currently, we allow live impersonations, and there are no dividends made off of these impersonators. But to create custom content, and using the person’s replicated voice seems like an invasion of privacy. Some celebrity stalkers feel like they are connected to a celebrity already just through watching videos and looking at pictures. This can eventually lead to criminal activity, such as breaking, and entering into the persons house (see Sandra Bullock’s B&E case here), kidnapping & robberies( see the case of Kim Kardashian here) or even murder (see the case of John Lennon here). How much more could this escalate if these stalkers are buying custom content? Will it increase the crime rate and jeopardize any person’s safety and security due to the desire for the stalker to feel even closer to the person? These are all interesting factors to consider.

AI Generated Voices, Videos & The Sex Industry

I would imagine that there may be ways to create custom content for those willing to pay. For example, some pornography companies may use AI to create videos of desired celebrities or even a crush the person may have in their every day life. These production companies may be able to make a script custom to the buyer’s desires, using the mimicked person’s person’s voice and some sexual acts that the viewer may be interested in. I’d also imagine this voice technology can also be used for sex dolls to impersonate the desired subject.

Taking it one step further, I also think this technology can be used in conjunction with virtual reality (VR), specifically for sex scenes or any type of interactions with people across the globe. It seems like the possibilities are endless.

What Does This Mean For the Future?

I was having an amazing conversation with a lovely gentleman yesterday and he said it seems the technology follows a trend where it exponentially changes every 10 years. He said the last 10 years was social media and he believes that this will fade away slowly as AI takes over. My response was that it is hard to imagine a world without social media because a lot of businesses depend on it for connectivity, advertising etc. I am curious if there will be something else coming used specifically for business marketing, similar to LinkedIn, but different.

I think the whole idea behind creating custom videos and voice content will be centred around giving access to people who aren’t accessible. The cost for accessibility to the human themselves, in the flesh, will increase exponentially. This can also mean through media like interviews, commercials and movies that are not AI generated. I’m sure there will be a market for creating content of the deceased, both known personally and known, based on celebrity status, to the viewer. I watched a Michael Jackson voice cover of a song released after his death. It gave me creepy vibes. I can’t imagine seeing my deceased relatives through AI. I’d imagine my brain would be unable to handle the confusion and it would probably tarnish my memories by attributing inaccurate personality traits to the character. However, some people may find comfort in bringing the dead back to life and may even be a useful tool through grief therapy, whether it’s through losing a person through death or some other means, such as a breakup.

I will keep daydreaming and writing about where the future will possibly take us. And I could be completely wrong but it is still fun to theorize. Regardless, we are in for an interesting ride in the next 10 years so buckle up.