Another day, another AI company caught using copyrighted works to train their AI systems without consent. Meta joins the growing list of artificial intelligence developers accused of stealing data from authors and artists, leading to a lawsuit.
Meta Faces Lawsuit for Stealing Data for AI
The social media giant joined the AI race back in February, adding or planning to integrate AI tools into its products. It looks like Meta is following in OpenAI and Google's footsteps as it is hit by a class action suit accusing it of stealing data for its LLaMa software.
One of the plaintiffs is a Pulitzer Prize-winning author, Michael Chabon, who also claimed that Meta infringed his intellectual property. The group of authors says that Meta used their texts to train its dataset, as reported by Gizmodo, all without credit, consent, or compensation.
Meta claims that the dataset used 85 GB of training data from a "books" category which came from Project Gutenberg, which is an online text archive with around 70,000 books that aren't copyrighted, along with a Books3 section of ThePile.
ThePile is where the legality of it starts to get murky. The lawsuit states that the books are compiled by a private tracker called Bibliotik, which is a shadow library found through torrent systems. These libraries are more often than not, illegal, as stated in the lawsuit.
Many of Chabon's content are on Books3, including those that are still under the protection of copyright. The same goes for other plaintiffs in the class action suit including David Henry Hwang, a Grammy Award-winning playwright.
Google President of Global Affairs Kent Walker says that it's similar to students going to the library and learning how to write and read, all while making sure that they are not reproducing other people's works and doing things that would vilate copyright.
Regardless of the logic or comparison, the works of the affected authors were still used without consent. As the lawsuit states, they are "entitled to statutory damages, actual damages, restitution of profits, and other remedies provided by law."
Microsoft Will Protect You Against Lawsuits
Creators such as authors and artists often go after the company that develops the AI systems that infringes their intellectual property, but users of the AI tools can also be hit with legal reprecussions when the tools generate something similar to licensed works.
This is especially true when the users who utilized the AI tools that created derivative content claim that content is theirs or if they profit from it. But, they can escape such legal actions when they use Microsoft's AI tools and products.
According to Ars Technica, the software giant will provide legal pretoection to its users when they are sued by creators claiming that their worked was infringed, as long as they use the built-in guardrails and content filters.
The company says that it will assume responsibility for the potential legal risks involved, as well as pay legal damages or settlements for the users of Copilot, Bing Chat, and other AI products under Microsoft.
