The music industry is watching the swift evolution of generative/creative AI technologies with a mixture of fascination and concern. That evolution is being backed by increasingly-large funding rounds for the biggest firms in the space, as well as for emerging startups.
There are now some figures on that, courtesy of data from investors’ database Pitchbook. It shows that generative AI startups raised $5.9bn of funding in 2022, up from $1.5bn in 2020.
“Ballooned” is the word Reuters used to describe the increase in its report. This covers all sub-categories of generative AI: models that can produce text, images, video, code and, yes, music.
While music will account for a small proportion of the $5.9bn raised in 2022, Music Ally has been tracking an uptick in the size of rounds raised by AI music startups in the last couple of years.
Funding rounds in the double-digit millions have included Splash ($20m in 2021); Aimi ($20m in 2021); Endel ($15m in 2022) and LifeScore (£11m in 2022, which was $14.5m at the time). Several earlier-stage startups have raised single-digit rounds, meanwhile.
Stability AI raised $10m in 2022, and while it’s best known for its Stable Diffusion text-to-image AI, it is also involved with music with its Harmonai subsidiary. Soon after that round was announced, it hired AI music veteran Ed Newton-Rex as VP product at Harmonai.
Funding for this sector is good, and not only because it will help these startups to hire more staff and improve their models. It also paves the way – perhaps – for licensing deals with music rightsholders to make their catalogues available to the startups for training purposes.
Earlier this month, we wrote about the launch of the Human Artistry Campaign, a group of creative industry bodies looking to set principles “to ensure artificial intelligence technologies are developed and used in ways that support human culture and artistry – and not ways that replace or erode it”.
“AI must be subject to free-market licensing for the use of works in the development and training of AI models,” was one of the principles set out by a group whose members include the IFPI, RIAA, BPI, NMPA and SoundExchange. “AI developers must ensure any content used for training purposes is approved and licensed from the copyright owner.”
Training cropped up as a key topic at the IFPI’s recent launch of its Global Music Report too. “The bottom line is, a lot of AI developers are just ignoring the ethics of ingesting the creative work of others, or they’re simply justifying it with what we view as a dangerous distortion of the idea of fair use,” said UMG’s digital boss Michael Nash in that report.
(The US trademark for the Human Artistry Campaign was filed by Universal Music Group, incidentally.)
Two thoughts to close. First: increasing funding for the generative AI sector, AI music startups included, only creates a path to licensing deals if the music industry figures out what those deals should look like. That’s important work.
Second: don’t be misled by the huge numbers floating around the wider creative-AI sector’s funding and projected future revenues. Musical AI is still one of the smallest chunks of both, and its startups are cutting their cloth accordingly.
The deals we bring to them will need to be flexible and workable, to build partnerships in an area whose future impacts are still so unclear. Our hope is that both sides come to the negotiating table in the right spirit: startups seeking meaningful licences, but also rightsholders offering them.