How people made their data licences
A few months ago, I wrote about making an exhibit for Big Bang Data at Somerset House that let people try a way we think licences could be made. After four months, 15,571 licences were created. Some of these licences were likely created by playful kids and over-excited adults, so we’ll treat this data with caution. Regardless, it’s interesting to see what’s behind it. Here’s what we found out.
Data licences are a design pattern for sharing data. They put people in control of their data by letting them set the terms of how it is shared.
Participants began creating their licence by scanning an object that represented a data type. These were: a travel card for location data, a smart thermostat for household data and a fitness tracker for biometric data.
Most people, 6,382, made a location data licence. The location object looks like London’s Oyster card, so we think familiarity caused its more frequent selection.
Levels of sharing
Data licences let people choose who their data is shared with. It can be kept private, or shared openly in the “data commons”, a speculative repository of data that allows free access to anyone, for any purpose.
Licences can also be customised by setting rules on who the data is shared with. Half of participants selected this option, suggesting a need for sharing data using terms that an individuals sets and understands.
Everyone vs. Non-profits only
People who agreed to share their data under set terms were asked three more questions to customise their data licence.
First, we asked if they would allow commercial use of their data or only with non-commercial users. 68% of participants chose to share for non-commercial purposes only. This includes organisations like research groups, foundations and charities.
It’s easy to conflate non-profits with being more ethical, but this isn’t necessarily true. Cutting off commercial use of data can be bad for the Internet economy. It’s something Sarah and I have written about in the past, particularly in online advertising.
Sharing with government
Next, we asked about government access to data. We asked people to choose between all government access, local government only or none at all.
61% of people chose to share their data with government, mostly with local government. It’s interesting to consider how this percentage may have been influenced by changes in public attitudes towards governments and personal data, particularly after the Snowden leaks about government surveillance.
Finally, we asked participants whether they would let their data be shared with third parties under the same terms, using the same data licence.
The results don’t show a distinct preference. While making the exhibition, we spent a lot of time on the phrasing of this question. Maybe we didn’t explain the benefits of data sharing clearly enough. It has great benefits in areas like drug development research, where many organisations may have to collaborate and share data.
When talking about data, it’s important to use cues from everyday life to make this conversation more accessible. Illustrating the three data types through physical objects helped people understand how data is used in their daily lives. We saw this when most people explored the uses of location data through the travel card object, representing what they had possibly used on the way to the exhibition.
People want to take control when they are offered it and when they understand. The exhibition showed that people want to share their data, but we’re confronted by impenetrable terms in every device or service. These rigidly set out how our data is shared and used. We almost always accept them, rather than miss out on what these things offer us.
In reality the way we want to engage our data with these things is more nuanced. Todays opt-in, opt-out system of terms doesn’t reflect our needs. What we made for this exhibition is a first try at making something that does.
Thanks to everyone who visited the exhibition. If you missed it, you can give the licence builder a try on our website.
You can also download the raw data. (JSON, 2.9 MB, CC-BY). Let us know what you find out at @projectsbyif.