Learning Robots Confront Copyright Law

Siyu Yan is an L.L.M. candidate, 2020 at NYU School of Law.

With the development of AI, many learning robots are emerging in China. A learning robot is a type of mid-size AI machine, with the appearance of cartoon characters. The machine incorporates different versions of textbooks on different subjects. The robots are able to read the content out, interact with children, and answer their questions about the content. Some are accompanied with functions that allow children to talk in English with the robots, and some can play music for children. There are also some machines that have the ability to act like a monitor camera for child safety when parents are away for work. Specifically, parents could talk to their children via the machine and control the content broadcasted to children.

The market for learning robots is still an immature one. The manufacturers usually incorporate various versions of textbooks that they are able to locate from the education market.

Copyright issues will play a significant role with regard to these learning robots. First, is the issue regarding the appearance of these robots. Many robots adopt the appearance of existing cartoon characters or make alterations based on existing characters. Of importance is whether this sort of imitation would infringe the copyright attached to these cartoon characters.

The second copyright-related issue concerns the music that is played from these learning robots. While some songs may reside in the public domain, a large majority of songs will likely be protected by valid copyrights. Those developing this technology will have to confront the issue that these learning robots may, in fact, infringe the copyrights associated with the music composition or sound recording work.

The third issue involves the textbooks incorporated in these machines. The manufacturers make great profits from consumers who care about the education of their children, and the activities of these manufactures are commercial. In their hard-copy form, many of these textbooks are published educational materials and are not free. Students have to pay for the physical copy. But with a learning robot, students will have access to these textbooks, free of charge. This scenario will confront not only the possibility of infringing the literal copyright in the textbooks, but also the performance rights that come with the copyrights.

Finally, an important legal conundrum will deal with how to approach the software (or code) used to operate these robots. As a threshold matter, it is an open question whether courts will grant copyright protection to this software. This question is important insofar as copyright protection will prevent competitors from freeriding off of the software already created. Related to the issues surrounding of copyright protection, these learning robots raise other interesting questions. Specifically, How could manufacturers prevent others from manufacturing competing robots with identical or similar functions? Which copyright claims could they raise? From a practical view, if litigation were to be brought against manufactures or among the competitors, the availability of evidence may be sparse. This is in part due to the fact that China does not have discovery procedures during civil litigation.