The U.S. Library of Congress adopted the U.S. Copyright Office’s recommended rule regarding the Digital Millennium Copyright Act (DMCA) application on AI (artificial intelligence) security researchers.

According to the decision, researchers seeking to examine whether AI generates prohibited content—such as discriminatory or explicit material—will not be exempt from the DMCA’s provisions that forbid bypassing digital protection mechanisms that limit access to copyrighted material. Consequently, circumventing technological measures such as encryption that protects works will remain illegal, even for these research activities.

This decision essentially denies exemption requests made by AI researchers. The Copyright Office stated that the requested exemption exceeds the scope of the DMCA and is a matter of policy that ought to be addressed by lawmakers.