Yesterday, this technology seemed so far away. Please excuse the Beatles-inspired pun. These days it seems like new applications of AI are making headlines everyday. In this post I'll share my opinions and some context regarding the recent usage of AI in The Beatles' latest and last ever song, Now and Then, and the ethical issues associated with this type of technology.
I think it's safe to assume The Beatles need no introduction. As probably the most influential pop music group of all time, their legacy has spanned generations. It's incredible to see that AI has enabled them to release new and unheard music 50 years after the band was disbanded.
The song has a long history, originally being written and sung by John Lennon before his tragic death in 1980. The track was later passed on to Paul McCartney and multiple attempts were made by the surviving Beatles members to complete the song, but the track had audio issues that could not be resolved until recently. New technology, developed by VFX studio Weta, has allowed Paul McCartney and Ringo Starr to complete the song, with contributions from both John Lennon and George Harrison.
The song is accompanied by a music video (see below), which includes new and old footage merged together to show all the Beatles members dancing and rocking out one last time. Whilst some find the video unnerving, I find it symbolic for what the technology has done, which is give us one last true collaboration of one of the greatest musical acts of all time.
For those curious about the technical side, the software that made this possible is called Mal (machine audio learning). It was originally developed for forensic investigation purposes, for the police to hear conversations in low quality audio. The software is not available to the public, but can supposedly split any music into separate tracks.
I don't think anyone has qualms with this usage of AI, as this audio is still Lennon's voice, just cleared up. There have however been other recent cases of AI in the music industry that have evoked ethical discussions. One recent example is a viral song that used AI to imitate the voices of popular artists Drake and The Weeknd. This song, titled 'Heart on My Sleeve' brought a lot of attention to the positive potential of this technology as well as its potential risks. Whilst ultimately this type of music may be appreciated by consumers, its debatable whether music enhanced by AI can be considered authentic if the technology has been trained on other artists' music. My opinion is that modern fans and consumers do and always will care about the image and the symbol of the musician as much as they care for the music, more now than ever in the age of social media. I don't think one can exist without the other and I don't believe music using AI for imitation will have any staying power.
As for the future of AI in the music industry, there will undoubtedly be more tools developed that will make music-making more accessible and potentially easier for everyone. Some artists seem to be leaning into the new innovations, with music artist Grimes even licensing their voice for usage in AI tools. Another question that arises is who should be compensated for AI-enhanced music, contenders include the original artists in the training data, the developers of the AI technology and the users of the AI technology. It is also somewhat frightening to imagine a future where high quality deepfake audio technology is widely accessible.
The potential to separate tracks from old audio is a great benefit, whilst the ability to imitate the voices of others opens up a range of ethical questions. Technology like this is usually hard to govern once it's out there, so it will be interesting to see whether the public will be open to consuming fully AI-generated music, and whether they will even be able to tell the difference.
How do we balance the preservation of artistic integrity with the possibilities offered by AI innovations?
Does music with AI vocals and production represent a new frontier in music, or is it simply a fad that will come and go?
Feel free to share your thoughts.
Links for further reading:
What a time to be AI 😎
Interesting read, and I think when you point to ai-use in music creation rather than text/literature, the point is made even clearer:
can you make true art without a “soul”? Who are we when we stop being able to tell the difference?
In this case it seems to be a tune Lennon&Harrison was involved in the making of (which in some ways justifies it) What if the next step is to develop totally new music in which an artificial brain claims to have replaced the genius style of these musicians? Should this be allowed to utilize commercially?
..an ethical question : would John Lennon and George Harrisson approve? Is it ethical OK to publish something in the name of someone that doesn’t exist anymore?