Reader in The City Law School, Dr Enrico Bonadio, is of the view that with machines learning to mimic human creativity, the copyright world has accordingly entered into AI-driven unchartered territory.
Speaking at a recent seminar on ‘Algorithmic art and copyright’ hosted by the Business School (formerly Cass)'s Centre for Creativity Enabled by AI, Reader in The City Law School, Dr Enrico Bonadio, stated that artificial intelligence can create art in its own right- refuting the notion that computers are merely tools for creating art.
Dr Bonadio says art, music and literature are increasingly created by machines and robots. He cites the $432,500 sale of the AI-created portrait Portrait of Edmond Belamy by Christie’s in October 2018 and Jukedeck, a program which employs machine learning to train algorithms that understand the rules of music theory and is able to create music.
“Customers can set parameters, such as music genre (e.g., jazz, blues, rock, ambient, chillout, etc.), main instruments, length of the tune, and its speed (e.g., up-tempo, low-tempo). The track is then composed in around 20 seconds, and made available to customers as an MP3 file; or the Cybernetic Poet, software which allows a computer to write poetry by adopting style and vocabularies of human poets (this technology is also protected by a US patent).”
Dr Bonadio says that “if such musical, literary and artistic expressions were created by humans, no one would object to them being considered as copyrighted works.”
They are certainly capable of captivating an audience and stimulating emotions and intellects in a similar fashion to works of music, literature or art produced by human beings. Yet, these outputs are generated in large part (and sometimes almost entirely) by machines – and the human being(s) who triggered the automated process that lead to the creation of the expression have often no idea how the ultimate work will look or sound. In other words, this ‘unforeseeability’ feature seems to break the causal link between humans who create or use these machines and the final output produced by the latter.
He goes on further:
“As machines have quickly learned to mimic human creativity, the copyright world has accordingly entered into AI-driven unchartered territory. Are and should AI-created works be protected by copyright? How do we answer the question of authorship (and ownership)? To what extent the use of data fed into the system, for example to train the algorithms, may amount to copyright infringement, or may in certain circumstances be exempted under fair use, fair dealing or similar doctrines? These issues must still find adequate answers – and the law does not seem prepared to address them yet.”