Scintilla – a flash, a spark, an iota. Shorthand for creativity and an indicator of inventiveness under Australian law.






Thursday, June 2, 2016

I, Robot. I, copyright owner?

By Emma Gorrie, Lawyer

A couple of years ago we mulled over copyright ownership issues when a monkey took a selfie. Another equally hairy copyright question is: who owns copyright in robot-generated prose?' More generally, can copyright even subsist in works generated by computer programs and, if so, who is the copyright owner?

To give some context, one of Google's latest artificial intelligence projects is trying to make better conversationalists out of gadgets. To this end, researchers inputted 12,000 e-books into a language model. The researchers then commanded the model to fill the gap between an opening and closing sentence to form a paragraph. If you are curious to know more, read the research paper (PDF). So does copyright subsist in the resulting robot writing? And if it does, who owns copyright?

Currently, the position in Australia is such that copyright simply does not subsist in works not authored by humans. The UK recognises copyright in computer-generated works, which is owned by the person who undertook the arrangements necessary for the creation of the computer-generated work. The EU recognises copyright in databases, identifying the right holder as the natural person who created the base. The position in each of Australia, the UK and Europe is discussed in more detail below.

Australian Law


Under s 35(2) of the Copyright Act 1968 (the Act), the 'author' of a literary work is the owner of any copyright subsisting in the work. As we reported in the Monkey Selfie post, an author of a photograph is defined in s 10(1) as the 'person' who took the photograph. There is no definition of 'author' in relation to a literary work and no general definition for 'author' in the Act. However, as stated in IceTV Pty Ltd v Nine Network Australia Pty Ltd at [98], the generally accepted position is that the author is the 'person who brings the copyright work into existence in its material form'. 'Person' is taken to mean 'human' and there is no provision for monkeys or robots. This arguably narrow view of who (or what) can be an author of a work is tied to the requirement of some 'independent intellectual effort' for a work to be considered 'original' so that copyright can be deemed to subsist.

You may be thinking that the author of a work generated by a computer program is clearly the person controlling the program. In the Google example, you would perhaps point to the researchers who inputted the e-books and ran the language model. However, this will only be the case where the efforts of the person controlling the program 'can be seen as directing or fashioning the material form of the work'. If the person controlling the program cannot be seen as directing or 'fashioning the material form of the work', then he or she is not contributing sufficient 'independent intellectual effort' to have 'authored' the work. These were the observations of Justice Perram in Telstra Corporation Ltd v Phone Directories Company Pty Ltd, where it was held that copyright did not subsist in phone directories compiled by automated processes from databases. Justice Perram gave the analogy that planes in auto-pilot mode are flying themselves.

To the extent that the activities of Google's researchers can be compared to a plane in auto-pilot mode, any arguments as to copyright subsisting in prose produced by the language model are likely to be short-circuited by the inability to tie the output to some 'independent intellectual effort' of a human. At least, this is the position under Australian law. To keep the analogy going, if the researchers were 'controlling the flight path' so to speak, copyright could be found to subsist.

The law in the UK and Europe


Section 9(3) of the Copyright, Designs and Patents Act 1988 (UK) deems the author of a computer-generated work to be the person 'by whom the arrangements necessary for the creation of the work are undertaken'. A 'computer-generated' work is defined as a work 'generated by computer in circumstances such that there is no human author of the work'.

The European Database Directive, adopted in 1996, recognised that databases could be protected by copyright where, by reason of the selection or arrangement of their contents, the database constituted the author's own intellectual creation. In addition, the Directive created an exclusive right for database creators to extract and/or re-utilise the whole or a substantial part of the contents of that database. 'Extraction' and 're-utilisation' include transfer of the contents of a database to another medium and making the contents of a database available to the public. This exclusive right applies irrespective of, and without prejudice to, protection of the database contents under copyright or other rights.

By virtue of s 9(3) of the Copyright, Designs and Patents Act 1988 (UK) and the European Database Directive, the fact that a non-human author produces a work is not fatal to the protection of that work under copyright and database law. This means that UK and European law effectively bypass the issues discussed above in relation to the Australian position.

Where does this leave the robot prose (and technology-generated works generally)?


There is a much better chance that the prose produced as part of Google's artificial intelligence project, and indeed works like it, will be afforded copyright protection in the UK or Europe than here in Australia.* Speaking more generally, if Australia seeks to reboot its approach to subsistence of copyright in relation to works produced by robots, it can look to the UK and Europe for guidance.


*Note the effect of the Berne Convention for the Protection of Literary and Artistic Works (1886), art 5. Under art 5, works that were not authored by a human will not be protected in Australia, irrespective of the country of origin of the work as defined. However, computer-generated works that originated in Australia will be protected in those Berne Convention countries that recognise the subsistence of copyright in computer-generated works.

No comments:

Post a Comment