Data Art: what are the limits and opportunities in data licensing for artists?

A session on using open data in artistic works of various sources, led by Leela Collins.

Traditionally, we have infographics, where we take data and visualise it so people can understand it. And then there’s conceptual art, which gains some of its meaning from the original data source. Does that create a new work, or does it owe something to the data producer?

Data is becoming a tool, in the same way that brushes are.

And then there’s protest art, where the whole of the data is used to create the art. But if the data is licensed non-commercially, can the artist make money from the work? A full open data licence is free for reuse. However, a non-commercial licence on some data is somewhat ambiguous – is it just restricting resale of the data itself, or does it prevent it being used for anything commercial?

Databases can be subject to copyright. But a collection of public facts can’t be copyrighted – you can’t own facts – but you can copyright the arrangement of those facts. Private datasets are more problematic, and you need to be careful with those, especially with protest art.

Some Creative Commons licences have a “no derivatives” clause – but they’re not really designed for data licensing.

Your work remains your own work, even if you have used open data. Copyright still applies. Artists spend years developing their style and approach – they bring a lot of themselves to their work.

If you’re considering licensing work – consider how comfortable you are with various uses – and then explicitly license based on that.

St Columba was responsible for the first copyright enforcement:

In 560 a dispute arose over a copy Columba had made of St. Finnian’s book of psalms. The result was Columba’s instigation of a rebellion by the Clan Neill against King Diarmait of Ireland. This culminated in the Battle of Cooldrevny in 561, at which three thousand men were killed.

Clarity and uncertainty in copyright

Being really clear up front helps so much: what are you expected to bring? What are you expected to give up? We’re dealing with issues where that clarity doesn’t exist yet.

People protesting Blizzard’s actions over Hong Kong have appropriated the character Mei from Overwatch for protest art. This is well-protected IP from a large corporation, yet people are getting away with it. There’s a lot of precedent for this – there are ways of doing it. But you need to protect yourself, your friends and your assets.

At least one attendee would be excited if an artists wanted to use her data in an art work. She, personally, would not expect financial reward for that.

The “share alike” problem

Open Data Camp 7 session on Data Art

Open Street Map is released under a share alike licence. If I create a painting from it, does that need to be “share alike” too? In theory, if you put enough work into that to make it a transformational work, then you don’t. But this is a difficult area and you might need an IP lawyer. In particular, if someone reproduced your work, would you be able to challenge that?

There are two issues here: scale and traceability. For many situations, it’s hard to connect the art work to the data source, unless you’re using tech that allows the connection – or until you work achieve a scale of commercial success that makes you worth suing. Much of this is untested in case law still.

Would a craft using a pattern based on personally-identifiable data in binary need to be GDPR-compliant?

One attendee has commissioned a project from a Unity engine developer to take any random data set and create a landscape you can explore from it, with a heads-up display giving information about “where” you are in the data. There’s significant creative work going that, making the decisions about how the data set becomes represented in the landscape.

Devolved government Hansard are licensed under the Open Government Licence, while the Westminster Parliament is licensed under the Open Parliamentary Licence. Obviously they have data about MPs in there – so there’s some conflict with GDPR, for example.

Newspapers have often created derived visual works from open data – but they never credited the data source. Should they? Yes, probably. This is just good manners. However – there can be a problem of scale, here, if you’re using lots of small snippets from multiple sources.