Overview
Below is this week's tracker of the latest legal and regulatory developments in the United States and in the EU. Sign up here to ensure you do not miss an update.
AI Litigation Update:
- Bloomberg filed to dismiss a lawsuit from Arkansas governor Mike Huckabee and other authors who claimed the company misused their books to train its large language model BloombergGPT. Bloomberg argues that the complaint did not specifically allege which authors' works were used, and is otherwise barred by fair use.
AI Intellectual Property Update:
- Per a blog post this week, the U.S. Copyright Office will issue its report on AI over the coming months in several sections: The first section, published this spring, will focus on digital replicas. The second section, to be published this summer, will address the copyrightability of works incorporating AI-generated material. Later sections will focus on the topic of training AI models on copyrighted works as well as any licensing considerations and liability issues.
- NTIA issued a report calling for establishing a system of audits for AI systems to increase transparency.
AI Policy Update—Federal:
- The Office of Management and Budget (OMB) sent out a guidance memo to all agencies regarding federal government use of AI.
AI Policy Update—European Union:
- A coalition of thirteen media and publishing associations, including the European Publishers Council, the Federation of European Publishers, and the European Grouping of Societies of Authors and Composers, among others, sent a letter to the European Commission's DG CONNECT, asking to be more involved in implementing the EU AI Act's rules on copyrighted material.
- The head of the Italian data protection authority (Garante) published an open letter directed to the Italian government and Parliament, underlining that the EU Member States' Data Protection Authorities are the best candidates to co-supervise and co-enforce compliance with the EU AI Act.
AI Policy Update – International:
China released its draft Artificial Intelligence Law. The draft law discusses how and whether AI providers can appropriately use copyrighted material for training purposes, coming out in favor of developing the industry over protecting copyrights. According to Article 24, AI providers can use copyrighted data in training sets, "if the use behavior is different from the original purpose or function of the data," and "does not unreasonably harm the legitimate rights and interests of the data rights holder." The draft adds that AI providers are responsible for "conspicuously" labeling the source of the data.
