I’ve covered AI many times here over the years. However, with debates and innovations continuously ramping up across the business, as well as the UK government looking more closely at potential legislation, I thought it might be a good time to examine some of the issues, as well as the innovations, in what is now an incredibly prevalent technology in many sectors.
In so many of the creative industries, AI has without a doubt opened the door for anyone to generate just about anything, and you need to trust me when I say that we are only at the very start of what is possible.
While some wider AI potential is absolutely terrifying (I’m mainly referring to weapons and political interference here), I’m genuinely excited to see how artificial intelligence develops over the coming years in other industries, including ours. However, my optimism is indeed still measured by knowing that our business is incredibly bad at agreeing on anything and that mixed messages going to the UK government right now could, once again, be our downfall.
Outside of productivity, marketing and the abundant benefits of aiding creativity with co-creation and assistive creation, there are various interesting and pressing subjects going forward in relation to AI in the music industry. These subjects include: the future market gatekeepers, the future of copyright, the licensing and records being kept of the works that are used, identifiers and labelling of human versus AI-generated tracks.
Then there is also the responsibility and accountability that DSPs hold on hosting deep fakes, the policing and monetisation of deep fakes, creators having a choice about whether their works can be ingested for learning (that ship might have sailed in some places, but absolutely must be addressed going forward), crediting in any and all forms, likeness legalities, monetisation and, finally, current and future consumption routes and usages.
Using AI-generated works for background music across various media types has always made a great deal of sense – and has existed in primitive non-AI forms for years – but the industry can easily imagine the potential losses which are at risk when the bigger fish across films, games, television (and any other sync revenue you can name) are replaced. It is happening already and is only going to grow. Furthermore, we are not the only creative industry that is concerned about this.
One particular part of the recent actors’ strikes included protesters asking for protection against their scanned likenesses being manipulated by AI without adequate compensation. Writers are also increasingly concerned about the market being flooded with AI stories and scripts. Indeed, back in February, one of the most respected publishers of science fiction short stories, Clarkesworld, halted applications after being overwhelmed by AI submissions.
I’m genuinely excited to see how artificial intelligence develops over the coming years
Then you’ve got AI versions of every artist popping up all over the place – which are, let’s be honest, really rather sonically impressive. These versions usually have hilarious AI names, but in the majority of cases they do not have any approval from anyone, and artists are currently, on the whole, neither credited nor paid [by AI platforms] for their voice, likeness or personality usage.
The AI music video platform Beatly popped up momentarily to host deep fakes, and claimed to do so without fear of legal reprisal, but the site has been throwing 451 errors for a long time now. When I messaged the creator, Alexander Zwerner, who is a student at Washington University in St Louis, he confirmed to me that the Recording Industry Association Of America (RIAA) forced his hosting platform to shut down the site.
And whilst DSPs have all scrambled to recognise and take down AI deep fakes, it’s worth noting that a quick search online will throw up thousands of them across various well-known video platforms, with advertisement revenue turned on.
On the other hand, you’ve also got products like Voice-Swap, founded by multi-platinum artist and tech founder Dan Stein, aka DJ Fresh. Voice-Swap, who work with law firm Clintons, claims to offer an SAAS platform for artists to monetise and license their AI voices with the protection of waivers, licences and watermarking.
Voice-Swap says that it offers 50% gross revenue share on its subscription model and 80% of licensing deals on AI vocals. They say they signed up 3,000 voice-swappers in two weeks when they launched.
It is clear that the industry and the government has no agreed policy to deal with all of this just yet, and I can’t help feel that the music business is forever caught unprepared with regard to new, and not always that new, tech. The horses have bolted yet again, and despite the bigger music entities being involved in AI for many years already, the industry has somehow, once again, failed to get protections or agreed guidelines in place to deal with it ahead of time.
There are important conversations happening at the moment, but it will take an aligned industry to create a way through that works. And alignment is not something we, as a business, are very good at, given that everyone tends to sit firmly on their own side of the fence in the middle of a drought, desperately trying to water their own garden. Ah, the music business, you have to love it. What a time to be alive, eh?