Lawsuits Filed Against Stable Diffusion And Midjourney
Mark Kennedy & Manuel Olumorin
January 23, 2023
GM. Lootbox here, grinding to bring the fountain of youth straight to your inbox so you don't have to.
Let's crack open the potions for the day.
- Midjourney/Stable Diffusion smacked with lawsuits
- Limit Break poses first viable creator royalty solution
- The Sneakies
Midjourney and Stable Diffusion Slapped With Lawsuits
Things escalated pretty quickly as 2 lawsuits were filed against generative art AI companies, including Stability AI (maker of Stable Diffusion) and Midjourney.
What are they?
- Class action lawsuit: Stability AI, Midjourney and Deviant Art are being sued by 3 individuals (Sarah Andersen, Kelly McKeran and Karla Ortiz) in a class action lawsuit in California for DMCA violation, Right of Publicity violation, Unlawful competition and breach of terms of service.
- Getty Images vs Stability AI: Getty images also brought a lawsuit against Stability AI in the UK for copyright infringement by scraping the internet for images and using them in its training dataset.
Both these lawsuits stem from these companies' use of the LAION 5B dataset, which contains copyrighted works of others that were used to train the algorithms.
Before the dread 😱 of losing our favourite AI tools set in, we went digging and found some fantastic work done by a lawyer on Youtube, Jake Watson at Corridor Crew, breaking down what the case is about.
Here's what we know now after going to Corridor Crew's School of Law
Many people think this case is about copyright infringement, and they are correct.
But it's bigger than that...
This case, when in court could determine the future of all AI tools, including our new friend ChaGPT, and the future of creators that have their works being used to train AI.
Before going into the details of this lawsuit, there are a few terms you should know.
- Fair use - A defence intended to prevent the rigid application of copyright law that would stifle the creativity that copyright law is to foster.
- Derivative work - Work derived from an original work with no change in meaning, expression or message.
- Transformative work - Work that alters the original work to the extent that it is unrecognizable from the original by using a new expression, meaning or message.
- Transformers - The algorithms that are used in these AI tools. They work by using encoders and decoders that add noise to the training images with text assigned to the image for context. The noise in the image is reduced using the decoder, creating a reproduced version of the image since it is not 100% perfect.
Back to the lawsuit
Since Stability AI is open-source, you would think this copyright infringement lawsuit is a nothing burger. But the open-source argument didn't work so well for Napster. So what's going on here?
- The lawyers filing the lawsuit argued that although these AI tools don't use the actual images to create new images (because of how transformers work), they retain latent images of the copyrighted work. Therefore, art generated from these tools are simply reproduced copies of the original.
- The strongest argument they make is works from the AI tools are derivative work, meaning they are infringing on artists' copyright since derivative work cannot be created without authorization from the original owner.
- Stability AI and others are yet to respond to the lawsuit, so they could settle or make a defence.
- If Stability AI did make a defence, their strongest argument would be that their use of the LAION 5B dataset falls under fair use.
For a fair use case, the courts consider these four things and precedence to determine that it is indeed fair use.
- Purpose and character of the use of the new work and if it is transformational enough that it is unrecognizable from the original work.
- Nature of copyrighted work.
- Amount and substantiality of the use of the original work.
- The effect on the market value of the original work, reducing or replacing its value.
Looking at these criteria, it is easy for Stability AI to argue that the work generated from its AI tool is transformative, and it does not necessarily prevent the artists from selling their original work. They also have two precedents that could work in their favour where judges ruled in favour of Google. The cases are Authors Guild v. Google and Perfect 10 v. Google.
Our take
Hard to see a world where Stability AI and others settle and not go in with a fair use defence. Definitely keeping an eye on this 👀.
Limit Break's Opt-In Royalties: NFT Creators and Collectors Win Big
Limit Break recently introduced opt-in, programmable royalties contracts for ERC-721 contracts through a novel staking solution. This is especially important for NFT creators, as it gives them more control over how royalties are distributed for their digital assets. The system is opt-in, so it's up to individual NFT collection holders to decide whether or not to "stake" and opt-in with their NFTs. This release will allow for numerous applications with the potential to reshape the NFT landscape radically.
It's a big win for creators. Here's why
- It allows creators to decide how royalties are distributed on the smart contract layer, giving them more control over their digital assets. This is a significant step towards empowering creators to have more autonomy and control over their creations. A value key principle of the web3 movement.
- The system is opt-in, meaning existing creators can adopt this solution without disrupting their existing collections. This is important for creators who want to retain control over the distribution and use of their digital assets without having to start from scratch.
- The on-chain royalties are enforceable at the individual token level, meaning each NFT can have its own programmable royalties. This gives creators more granular control over how their assets are used and monetized, which can lead to more equitable and sustainable revenue streams.
- The staking solution is backwards-compatible, so existing creators can adopt it without disrupting their existing collections. This means that creators can continue to sell and trade their NFTs as they usually would while also taking advantage of the programmable royalties feature.
It's also worth noting that Limit Break has prioritized two key objectives aligning with Web3 ethos: backwards compatibility and freedom of choice. This is why they made sure that this is an opt-in staking solution, allowing existing creators to adopt their solutions while letting their collectors decide whether to opt-in on an individual NFT level.
Our take
Overall, we think it is a great thing that Limit Break is innovating in the area of creator royalties for NFTs. Of course, as with any new technology, it's essential to keep an eye on the fine print and be aware of any potential risks before diving in headfirst. But overall, this is a step in the right direction for creators and collectors alike.
The Sneakies
- Blockchain Game Alliance Survey reveals industry professionals predict gameplay improvements to drive blockchain gaming adoption in 2023
- Oasys unveils Decentralized Governance for blockchain gaming.
- Game developers remain skeptical of metaverse and blockchain projects in GDC's new Survey.
- Scuti partners with Sky City's metaverse, Alterverse, to a new virtual shopping experience that makes any object or billboard interactive and shoppable.
Love what you just read?
Forward to a friend or share it on social media using the social media buttons at the top of this email.
What do you think of today's edition? |