Will AI Help Or Haunt The Creative Industries?

26 Apr 2024

Artificial Intelligence is developing at an alarming rate, or that’s what the fearmongering headline screech. As it stands we’ve got a fad development that’s more made fun of than made useful.

Since its inception in the 1940s and ever since DeepBlue became the first machine learning software to beat a chess world champion in 1997, the mainstream has been increasingly concerned about the potential consequences of AI. Despite murmurings from the scientific community that it’ll never achieve world domination and that it’ll only be as successful as the programmers behind it, there is still a fear.

Today, it is easy to side with those weary of AI’s potential considering the hike of AI generated covers doing the rounds on TikTok. With #aicovers raking in 12.2b views, all users have seen gimmicky covers of popular songs sung by iconic cartoon characters or Frank Sinatra’s voice uncannily singing Coolio’s Gangsta Paradise.    

Although, we are in no position to meaningfully comment on the longevity of AI in music as it is such an emerging field, of which we’ve barely passed the threshold. There are comparisons to be made to similar innovations, the recording industry frequently sees new concepts once scrutinised become second nature to engineers – such as autotune or even programmed instruments.

We are currently seeing two distinct forms of AI being implemented in music; AI generation and AI enhancement, enhancement being the only we’re seeing being used as much more than a fad. 

Take The Beatles’ ‘final single,’ Now And Then, originally a home demo recorded by John Lennon in 1977 restored and enhanced by machine learning AI with the surviving band then adding to the vocals – a project spearhead by Peter Jackson. Yes, that Peter Jackson director of the Lord of the Rings movies.

It is easy to forget what AI really is and to let the hypotheticals become our expectation. AI is essentially a programme expected to learn from exposure to excel beyond its base model. What was done with the Lennon demo was really just an enhanced version of audio isolation, using software coded to progressively improve.

Google have also dipped their toes into the AI Generation game recently unveiling Lyria, a software intended to generate vocals, lyrics and backing tracks in the style of popular artists. This sees the creative project shortened down to the entering of briefs and, poof, you have a track. Currently only limited to 30 second clips in the style of ten artists, including Charlie Puth, Sia and T-Pain but expected to expand and obviously having the technical support of the search engine titan behind it. 

Lyria, or DreamTrack is touted as a means for songwriters to spice up the creative process, immediately hear their ideas and provide greater inspiration; against a backdrop of vows that AI will have a positive impact on music, it really only seems like a tool that’ll kill the inherently human aspects of songwriting. Fears that AI will take your job have always been expressed half jokingly in the creative industries as unlike manufacturing work or accounting, the arts have always required passion that a computer software could never replicate. However, we seem to be edging ever closer.

Plankton (Spongbob) – Beggin’ 1.2M likes, 12.5M views

Frank Sinatra – Gangsta Paradise -AI Generated

Full Version

Lyria Announcement