The decision by a U.S. court to continue deliberating the major lawsuit filed by several visual artists against Generative Artificial Intelligence platforms could call into question how these platforms can operate without infringing on the copyright of works that were used to “train” them.
The lawsuit
About a year and a half ago, several visual artists filed a lawsuit against AI platforms, including Midjourney and Stability AI.
The plaintiffs are claiming that those platforms copied their works in violation of their copyright and used them to train their AI-based image-generation systems and build the infrastructure needed to produce products.
The lawsuit was filed in a district court in California alongside several other legal proceedings initiated by copyright holders against AI platforms.
The court has now ruled that the plaintiffs’ copyright infringement allegations against the platform owners in this lawsuit are sufficient for the legal process to proceed to the document disclosure stage.
The significance of this ruling
The ruling and the continuation of the legal process underscore the need for statutory regulation of AI platforms, which store enormous datasets and are amassing significant power, particularly in relation to copyright law.
The AI platforms’ main defense is that their use of copyrighted works is sheltered under the “Fair Use” doctrine in copyright law and therefore, even if they copied the copyrighted works for the purpose of training the platforms, this does not suffice to impose legal liability on them in respect of copyright infringement. The court’s examination of this affirmative defense argument is likely to be a crucial question affecting copyright infringement cases against AI platforms in the future.
On a practical level, the court’s ruling incentivizes companies relying on AI-based technology suppliers to conduct risk assessments and ensure that every contract with such suppliers includes commercial protections, such as indemnity mechanisms, to protect their information and rights, as well as of those of their customers.
What is the situation in Israel regarding copyright and AI platforms?
At the end of 2022, the Israeli Ministry of Justice published a position that partially addressed this issue. According to the ministry, the use of copyrighted content for machine-learning is generally covered by legally permitted use arrangements and therefore does not constitute copyright infringement. This position is consistent with the aspiration to not “lag behind” in AI system development and to incentivize desirable competitive activities in this evolving industry.
We emphasize that this position has not yet been examined by the courts. Israel is only just beginning to contend with the legal challenges posed by artificial intelligence.
[View source.]