New initiatives from Facebook to provide scholars with access to data and new tools to give users greater context when they see advertising show a shift in policy at the company. In this blog, co-founder of Who Targets Me and Visiting Fellow in the Department of Media and Communications at LSE, Louis Knight-Webb, outlines the key areas on which he would like Facebook to collaborate with researchers and regulators in order to improve election transparency.
Applying transparency to Facebook data
Being able to consent to our personal data being used by third parties requires knowledge of what exactly we’re consenting to. At the moment it’s not clear how or where the personal data that goes into targeted advertising is collected.
My colleague and Who Targets Me co-founder Sam Jeffers recently downloaded his Facebook data. The first thing he noticed was that a dozen or so restaurants had used ‘custom audiences’ to target him with advertising. Despite having never visited these restaurants, somehow they had obtained and uploaded his email address to Facebook, thus placing him in a ‘custom audience’ and allowing the restaurants to target him directly with advertising. If only there was a way to know how they had obtained his information.
Tracking materials as they flow from producer through processors and onto the supplier is no new concept. As an example from the non-digital world, in the food industry, supply chain transparency means being able to trace the origin of the ingredients in a food product: ultimately, being able to tell which farm produced the grain that fed the pig that ended up in a supermarket oven pepperoni pizza. Could we do the same for Facebook data?
When Sam next sees an advert and questions how the advertiser got hold of his personal data, I’d like him to be able to see where and when he gave the advertiser consent to use his email address to target him with advertising. Facebook could collect this information from the advertiser when they create the custom audience, and display it to the user in the ‘Why am I seeing this?’ dialogue. Advertisers using custom audiences should have to prove they have permission to use the data from that audience, before Facebook allows them to purchase advertising. Pages that cannot prove they have permission to use custom audience data should be banned from purchasing advertising.
In the wake of revelations about Cambridge Analytica, accused of obtaining Facebook user data in an attempt to aid their psychometrically driven political campaigning, I put it to Facebook that having a total understanding of how our data flows from the moment we consent to our data being used, through to the moment we see a targeted advert, is essential to fair and transparent elections.
Improving how regulators access political advertising
The UK Electoral Commission (which oversees elections and regulates political finance in the UK) is largely in the dark when political advertising is purchased on Facebook. If a political candidate in an election forgets to file their Facebook advertising receipts to the Electoral Commission as part of the spending and donation return all candidates are required to submit, the regulator will be both unaware and unable to find out for themselves whether the candidate had purchased such adverts.
The same is true for politically motivated companies and individuals. Anyone can purchase advertising on Facebook, and only Facebook knows the extent of that person’s spend. If the amount spent goes over the non-party campaign limits, there’s no guarantee that anybody will notice, because that data is kept private between Facebook and the advertiser.
Furthermore, there is no way the Electoral Commission can validate whether a political party is sticking to the spending rules. The current rules state that a political party’s budget cannot be spent promoting a candidate, but how would anyone know if it were? Spending receipts contain next to no information on which specific campaign or advert the receipt relates to.
I believe it’s up to the Electoral Commission to engage with Facebook over this matter. Facebook operates globally and would have to engage every regulator so as not to be seen pandering to one government over another. Does the regulator have the capacity to check every single advert? Probably not. But they could operate a policy of real time random ‘drug testing’ during elections: ‘Stopping and searching’ every one in ten campaigns to ensure compliance.
Labelling political advertising
Facebook has revealed mockups of how political advertising on the platform will be labelled and visible to all in the future. This is a great move, albeit in response to a great deal of pressure from regulators and the media. We at Who Targets Me would like to see more cooperation between Facebook, researchers, the regulators and the public when new transparency features are introduced. I’d like to see features that do not compromise on transparency for users over the interests of the businesses and campaigns that use Facebook. Facebook should collaborate with the outside world to test different designs for transparency, and we as a society can then decide whether these tools go far enough to protect our democracy.
I would like to see more interfaces to Facebook data designed with regulators and researchers in mind. For example, new Application Programming Interfaces (APIs) that allow programmatic access to advertising in real time and at scale would allow the community to build tools that consume this data to solve local problems. This can only reduce the burden on Facebook’s moderation team.
This article gives the views of the author and does not represent the position of the LSE Media Policy Project blog, nor of the London School of Economics and Political Science.