Have any suggestions for the 3 Count? Let me know via Bluesky, Mastodon or LinkedIn.
1: Publishers Argue AI Firms Should Pay for Content as Copyright Case Advances
First off today, Jon Reed at CNet reports that The New York Times and other publishers have filed a brief in their ongoing copyright infringement lawsuit against OpenAI and have said that properly licensing news publishers’ content would amount to a “rounding error” in the financial picture of AI firms.
The publishers sued OpenAi alleging that the company unlawfully used their content to train its AI systems. Discovery in the case is winding down, though both sides have been filing various motions and briefs.
In the most recent brief, publishers argue that AI firms have diluted the value of their work with “pink slime” content. Meanwhile, OpenAI argues that, if publishers don’t want to be trained on, they should use established technical measures, like robots.txt, to prevent AI bots from accessing the content.
2: Saudi Arabia’s New Copyright Regulations: What AI Companies Should Know
Next up today, Latham & Watkins at their blog report that Saudi Arabia has implemented new copyright regulations, which provide exemptions for the development of artificial intelligence and a compulsory license regime for Arabic translations.
On the AI front, Saudi Arabia’s new copyright law carves out an exemption for the use of copyrighted works in the training of AI models. However, to qualify for this exemption, companies must comply with a detailed set of conditions including using only what is necessary, keeping track of the data used and minimizing harm to the original creators.
In addition to the new AI rules, the Saudi government has also introduced a new compulsory licensing regime for Arabic translations. Under this regime, Saudi nationals or entities may apply for a license to translate a copyright-protected work into Arabic if the original rights holder has not done so within a specified period, usually three years.
3: Rightsholders Can’t Use OpenAI and Anthropic to Dismantle Meta’s Seeding Defense
Finally today, Ernesto Van der Sar at TorrentFreak writes that Meta is facing a series of copyright infringement lawsuits over how it trained its AI model. However, as part of those cases, the company is making the argument that, seeding BitTorrent files is an inevitable part of using BitTorrent to obtain copyrighted material and, as such, is not an infringement.
In earlier cases, judges have ruled that AI training itself is not a copyright infringement, but found that the use of pirated material to train the model was. This argument prompted the $1.5 billion settlement between Anthropic and authors.
However, Meta is now arguing that the use of BitTorrent means that the resharing of pirated content is an inevitable part of the practice. Rights holds had asked the court to inquire how rivals, such as OpenAI and Anthropic, handled this issue, but the two parties have refused to provide information on their BitTorrent practices and the judge has said that they cannot force them to do so.
The 3 Count Logo was created by Justin Goff and is licensed under a Creative Commons Attribution License.
Want to Reuse or Republish this Content?
If you want to feature this article in your site, classroom or elsewhere, just let us know! We usually grant permission within 24 hours.