A quick guide to intellectual property in digital assets – metaverse, NFTs and generative (AI) art

For those of you who don't know me, I work part-time as an intellectual property lawyer (copyright, trade marks, patents etc) in London. I have published numerous articles on the intersection between IP and NFTs, the metaverse and AI through my firm, Herbert Smith Freehills LLP, and I thought it might be interesting to share some high-level takeaways. You can take a look at the firm's IP in NFTs guide here.

NFTs and the metaverse

If you're still catching up on the world of non-fungible tokens (NFTs) or feel like a deeper dive into digital assets and the metaverse, I can recommend starting with our comprehensive guide here. I also did a deep-dive into IP and NFTs in another blog post on this page, which answers questions like – what do I actually buy with my NFT.

In my legal practice, we have seen a lot of interest in NFTs and the metaverse from clients who are actively participating in the space – I know we are in a so-called 'NFT winter' at the moment, but there's still a lot of forward-looking interest from clients about engaging in metaverse projects, and the involvement of NFTs both as art and functional assets in the digital world. Equally as important, however, is the defensive interest in digital assets.

Anyone with an internet connection can create their own NFT collection, which could involve stealing images off the internet (which happened in the recent Yuga Labs v Ripps litigation – see our summaries here and here) or creating original content that may violate the rights of well-known brands (as happened in the Hermes v Rothschild litigation – see our summary here). If this happens, you may be forced into action even if you were never intending to be involved in the digital space.

The Ryder Ripps Bored Ape Yacht Club (BAYC) collection consisted of NFTs that linked to the original BAYC collection originally released by Yuga Labs. This worked because NFTs generally don't contain the digital asset themselves, but have to point to a website that stores the assets. So Ryder Ripps could just link his imitation / satirical collection to the same images. The Court found they sold around US$1.6 million worth of NFTs from the infringing collection.

From an artistic perspective, we have seen many successful artists release their own NFT collections over recent years. My personal favourite was Urs Fischer's CHAOS collection, released in collaboration with Pace Gallery (www.chaosoahchaos.com). Damien Hirst released a number of NFT projects, most infamously his 'The Currency' series (which I have previously discussed here and here), with each NFT being either solely a digital asset – or also being linked with a physical work. In addition, there has been the rise of digital artists who have become successful in their own right because of the advent of NFTs. Beeple, who single-handedly fuelled the institutional interest in NFTs, is one example. Other less well-known artists include Dmitri Cherniak and Tyler Hobbs. Earlier this year, Sotheby's were tasked with liquidating the NFT collection of the defunct Three Arrows Capital, with the highlights selling for US$17 million across two auctions (and one Cherniak NFT reached US$6.2 million). Remember, this was well into the NFT winter when prices in general (and of cryptocurrencies) had crashed.

AI

Something I haven't discussed before is the impact with AI. A number of cases relating to the use of IP protected materials in the training of AI engines are now before the courts of both the US and the UK (such as one brought by Getty Images against Stability AI, the maker of Stable Diffusion). These cases could have significant implications on whether and how businesses can protect their material which is publicly available and which may have already been used in the training of AI engines.

Another significant area of uncertainty for IP and AI is whether the output of AI engines can be protected under IP law in the same way that more traditionally human authored materials. The US copyright office has maintained its view that images produced by AI engines are not protectable by copyright due to the lack of a human author. This was confirmed by the US court in relation to an application made for protection for copyright images (see our blog post here). Whilst the protectability elsewhere in the world under copyright is still not formally determined, the US approach is a warning to companies utilising AI engines to produce output, and artists wanting to rely on that output, which may not have all the traditional IP protections that usually apply.

Finally, on the issue of whether the output of AI engines may infringe existing IP rights, companies have taken matters into their own hands to encourage the uptake of the AI products. Microsoft, Google and Adobe are all now offering some form of IP indemnity to give customers comfort that use of their respective AI products will not lead to liability for copyright infringement claims. For further discussion of the issues around IP and AI, we have another series here.

Please let me know if you have any questions, or there are any topics you'd be keen to hear more about!

George