Following on our recent blog article Getty Images v Stability AI – Key points for AI developers and users, there has been another notable settlement involving the use of original works by AI developers, involving Warner Music Group (“WMG”) and Suno.
In the US, WMG (alongside other major labels including Universal Music Group (“Universal”) and Sony Music) sued Suno for, allegedly, training its AI music-generation models on copyrighted recordings without permission and using the AI model to create new similar music. Although Suno rejected WMG’s claims of copyright infringement and argued that the use of the recordings constituted fair use under U.S. copyright law, the question did not need to be decided by the courts as the parties settled before judgment could be made.
This is not the first case WMG has launched against an AI company for copyright infringement of their recordings. Less than a month ago WMG and Universal settled their lawsuit against Udio which resolved a copyright infringement dispute relating to unauthorised use of copyrighted recordings to train Udio’s AI models. In the Udio case, the settlement involved a deal where WMG and Udio would collaborate to launch a new platform which would enable Udio to continue training its AI models on WMG’s recordings, through a licence and with artists who consent to their voices being used. Udio will have to implement protections to ensure that the music on its service will be authorised and licenced.
WMG also signed a deal with Stability AI, following Universal’s deal with Stability AI a month earlier, in which both parties referred to the use of ethically trained models and the advance of “responsible AI in music creation”.
WMG’s CEO, Robert Kyncl, lists his AI deal non-negotiable principles as:
These recent settlements and commercial ventures show that content owners are proactively protecting their intellectual property (IP), with a willingness to reach collaborative deals with AI developers, rather than continuing with litigation, and likely motivated by the risk of a decision which may go against the content owner, as we saw with the Getty Images case. The move towards collective consent and licencing must be seen as a positive for both IP owners wanting to protect their IP and AI developers ensuring that their use of an owner’s IP is non-infringing and legally compliant.
As the English courts have yet to decide a case on generative AI training and primary IP infringement, and while we wait for the UK Government’s response to its Copyright and Artificial Intelligence consultation, the fact that these deals are being done by major corporations in the US is likely to set the agenda of what will transpire in the UK.
Navigating deals in the 'AI era' involves a good understanding of the technology industry and intellectual property rights. At Roxburgh Milkins we have a number of clients within the above sectors and can help you structure these deals and negotiate the contracts.
If you are looking to protect your IP, collaborate with an AI business or if you are an AI developer, or if you just want a chat about how this might impact your business, please get in touch at kimberlee.carstensen@roxburghmilkins.com.