Extracting what’s relevant for my musician client base from the deluge of information on this topic.
As a tax advisor specialising in advice to musicians, I have been closely following the conversation around Generative AI in music. Not just because I love a clever app, or I am susceptible to #shinynewthing syndrome, but because of the potentially seismic shift that AI in music will inevitably bring. This shift is bound to affect my musician clients and eventually will trickle down to me in the form of income that needs to be taxed. So, it made sense to "follow the money" in order to understand how it was derived.
From what I've learned so far, there are several distinct stakeholders involved in the new world of Generative AI in music. Sometimes their interests align, and sometimes they don't. To make sense of this, I wanted to take a look at my own client base as a representative group and delve into their perspectives in greater detail...
Session Musicians and Generative AI
The first group that comes to mind are session musicians. For these performing artists, the conversation around copyright becomes moot as they are "bought out" of their copyright at the point of their participation in the recording.
This means they have no vested interest in whether the catalogue they feature on is exploited by AI or sold on in the future. Their rights were paid for upfront. However, this doesn't mean the AI story ends here for session musicians. There are aspects that they ought to consider for the future...
Session musicians should ensure their future buyout agreements recognise the potential sources of income that may be derived from their next bought out performance. This could mean pushing their unions and negotiating bodies to ask for higher buyout rights.
This is because the ultimate rightsholder(s) are now looking at a wide variety of new business models which will exploit their talent and participation on the next track. This includes the possibility of large record labels proactively licencing their material to music creation apps. Universal Music, Google In Talks To Develop AI Music Tool: Report (digitalmusicnews.com)
Composers and Generative AI
So this group are profoundly affected by this technological advancement. They find themselves in an interesting position which straddles that of the musicians discussed above and the ultimate rights holders. The main concerns that I've heard from composers revolve around two issues...
Firstly, the fear of AI taking over their job. This concern can be broken down further: the practical aspect where AI technology is making it so easy to generate music that composers can feel threatened that their role is becoming redundant.
Positives first: for amateur musicians, it democratises access to creativity. To enable someone who wouldn’t otherwise be able to express themselves seems to me to be a good thing…...My 14 year old-drum-kit-playing-TikTok posting son is a good example here. On the other hand, are kids going to realise there is a listening experience that can be so much more infused with subtlety than the output they are hearing? In other words, will the AI-generated music they get used to lower the benchmark for quality?
Do they even care?
For professional music makers, however, it's generally agreed that as of now, AI technologies are a "tool" and not a replacement for creativity. They can be used to enhance and streamline the creative process but can't replace it.
One of the sentiments I found comforting here was a comment made by @Valerio Velardo on the Music Industry Coffee Break podcast , and I’m paraphrasing here;
“a computer can’t feel; they can be told to produce something at a certain number of beats per minute, or using certain instruments, etc. but they can’t be trained to infuse the output with something as ephemeral as emotion.”
The second concern composers have, relates to the regulation of creative AI technology and fair remuneration. At present, a composer has to explicitly opt out if they don’t want their music to be used to train AI models. It is probably naive to believe the horse hasn’t bolted here insofar as big tech companies have already scraped your musical data up to now, but what can you do in the future? Composers should check their contracts with publishers and ensure that they do not give permission for their music to be included without consultation with them.
If they want to opt out of any usage of their material to train AI data models, they’ll need to get in touch with their collecting society and notify them of this. At the moment, the default assumption is that you are opting in.
Voice Artists and Generative AI
The third group of stakeholders are voice artists. This group is in a unique position because, and this is very simplistic analysis of an incredibly complicated area of intellectual property law but for now, suffice to say, the sound of a human voice is inadequately protected under current UK copyright law. There are already changes afoot here in the UK; in the EU and in the US so hopefully a greater measure of protection will be brought it but for now, that’s where things stand.
Unsurprisingly, there are pros and cons to what’s happening here too. One really big concern is that the lack of regulation means that voice-cloning and deep-fake technology will leave an individual artist vulnerable to being wrongly aligned with, say, a political viewpoint they might not agree with.
On the positive side, the current lack of regulation is beginning to worry some of the big consumers of this kind of product. Many companies that might previously have enjoyed the cost savings attached to content that didn’t incur licencing fees are now getting more cautious. They don’t want be seen to be exploiting musicians and voice artists and using content that is not traceable back to its rightful owners is too risky.
As a result, my findings are that many of the startup apps using voice cloning technology are presenting themselves as sensitive to the political pressure to do right by their co-creators. They appear to want to properly license the artist voices they are using as raw material. This means that singers and voice artists are in a position to exploit a potential new income stream by offering their fully licensed voice to these apps.
Obviously you need to be really careful to check the small print and make sure you are not allowing your voice to potentially be used for malicious intent or something you disagree with in the future but I think it’s an area worth exploring for professional voice users.
As a former professional singer, I reckon I’m as good a guinea pig as any on this one so I’m currently doing the rounds of these apps to see what the experience looks like and whether there actually is any rights based income to be earned here. I’ll report back on this in due course.
I think a lot of us are feeling overwhelmed with the constant AI news and updates, I know I have found it difficult to keep up. But two sources of information have been particularly useful for me in gaining understanding and knowledge. They are Music Ally and The Ivors Academy. So my thanks to both for their excellent coverage of these issues for musicians.
In conclusion, the emergence of Generative AI in music heralds a new era for the industry. I hope to see professional musicians exploiting the inevitable changes for the good of their industry as a whole.
Have questions on this? Contact us today.