Saudi Arabia Enacts Copyright Law Permitting AI Training on Protected Works

Saudi Arabia has implemented a copyright law that explicitly permits artificial intelligence developers to use copyrighted works for model training, EnterpriseAM reported. The Kingdom joins Japan, Singapore, and the European Union in adopting specific statutory copyright carveouts for AI data ingestion.

Under the law, which took effect in August, developers in Saudi Arabia may train AI models using copyrighted works provided the material was lawfully obtained and is used solely for training purposes. Developers are required to record the type, source, purpose, and date of each use, and trained models are prohibited from reproducing the underlying copyrighted works.

The law's implementing regulations also establish a mechanism enabling Saudi entities to translate works from foreign languages into Arabic without explicit author consent if no authorized Arabic translation has been published within three years or another designated period. While not designed solely for AI, analysts cited by EnterpriseAM note the measure could facilitate large-scale generation of Arabic training content across regional dialects.

EnterpriseAM noted the legal framework comes as Saudi Arabia seeks to expand its domestic AI model ecosystem and compute export capabilities, competing directly with the UAE to establish regional leadership in Arabic-language AI tools. State-backed AI entity Humain did not respond to requests for comment regarding the legislation.