Autocomplete suggestions: Did you mean 'copyright infringement'?

4 mins 09.01.2019 Nadia Braad, Ruby Ramachandran

The rise of the internet brings with it new ways of infringing the law and poses challenges to those defending their intellectual property rights. A recent High Court case may have relevant implications for copyright infringement actions.

Key takeouts

For an autocomplete suggestion to infringe copyright in a literary work, the work must receive protection under the Copyright Act and it must be a material reproduction of that work.

It may be a challenge to establish that the autocomplete suggestion is a reproduction of a literary work that is substantial enough to receive copyright protection.

It may be arguable that an autocomplete suggestion is a reproduction of part of a larger protected work, but this raises its own hurdles as to materiality.

While autocomplete suggestions can be defamatory, they are less likely to be copyright infringement because the words copied don't meet the threshold for being a literary work and are not a material reproduction of a work.

The High Court case, Trkulja v Google LLC [2018] HCA 25, considered whether Google's autocomplete suggestions could be defamatory publications. Relevantly for the purposes of this article, the Court stated:

… it is strongly arguable that Google's intentional participation in the communication of the allegedly defamatory results to Google search engine users supports a finding that Google published the allegedly defamatory results.

As the judgment dealt with an application to set aside the defamation proceedings, no final conclusions were reached.

If autocomplete suggestions are capable of being defamatory, then could autocomplete suggestions also constitute copyright infringement? To constitute copyright infringement, the autocomplete suggestions would need to be a literary work, the rights to which have been infringed by an online search engine. These rights might include the right to reproduce the work in a material form, the right to publish the work or the right to communicate the work to the public. Section 29 of the Copyright Act provides that a literary work will be deemed to have been published if and only if reproductions of the work have been supplied to the public, whether by sale or otherwise. This definition is distinct to the definition of publication in the context of defamation law, which requires only that the defamatory material has been communicated to some person other than the plaintiff. Regardless, given what the High Court said in Trkulja v Google LLC, and given that autocomplete suggestions are made to the users of the search engine, it seems likely that the autocomplete suggestions will be considered as having been reproduced or published by the online search engine.

The key question is identifying what a 'work' is. The starting point in defining a literary work is that it must be a work that was intended to afford information and instruction or pleasure in the form of literary enjoyment. Relevant additional factors include the originality of the work and the extent of skill and labour involved in its composition. There have been cases in which a Court has found specific titles and headlines were not works, because they were too insubstantial and short or because they were a means of referring to the underlying work. However, there is no hard and fast rule that a title cannot be a work, if the title or headline has sufficient substance. For example, Courts have suggested that there may be copyright in a whole page of title. Regardless, given that autocomplete suggestions are usually a short phrase, it seems unlikely that they would be considered substantial enough to be a work for the purposes of the Copyright Act.

Nevertheless, the ambiguity means that using a single sentence without the author's permission could end up in allegations of copyright infringement. In the USA, for example, a dispute arose between Twitter user and author of the “Why be racist, sexist, homophobic or transphobic when you could just be quiet?” tweet and Green Box Shop, which used the tweet on a T-shirt without the author’s permission. It is reported that the parties settled the matter without court proceedings, with Green Box Shop providing compensation and credit.

'That's not a knife' – a copyright infringement?

An alternative might be that the autocomplete suggestion is actually an excerpt from a larger piece that is considered a work for the purposes of the Copyright Act. A reproduction must, pursuant to s 14 of the Copyright Act, be of a substantial part of the work. While the entire work might be considered a literary work deserving of copyright protection, it is unlikely that a quote of a phrase from the work is a substantial reproduction. For example, typing into Google 'that's not a knife' will produce the autocomplete suggestion of the famous quote from Crocodile Dundee (not produced here for reasons set out below). While Crocodile Dundee receives copyright protection, it seems unlikely that the reproduction of a single line from the movie would be sufficiently substantial to constitute an infringement.

It's uncertain then whether the single line would be considered a material reproduction, but at least one person thinks otherwise. Paul Hogan began copyright infringement proceedings against the burger chain, Grill'd, over their use of the 'that's not a knife' quote on their knife covers. The proceedings ended without a trial in February 2018, so there will be no judicial pronouncements on whether the quote deserves copyright protection.

We won't be likely to see a Court consider this particular topic until the value of the copyright is sufficient for a copyright owner to challenge a search engine over the autocomplete suggestions. In the fast paced world of technology, this may not be far away.