This post was written Emily Sharpe, Web Foundation Director of Policy.
Social media platforms have a profound impact on democracy and elections. Narrowly targeted online advertising is now perhaps the most powerful tool that campaigns have to persuade voters and win elections.
To uphold the integrity of democratic elections, all social media platforms must ensure their tools are designed to support free and fair political processes.
As the world’s biggest social network, Facebook has an outsized influence in election campaigns and voter decisions. And so we’re deeply concerned that the company’s updated political advertising policy falls short of the standards required from Facebook to help ensure the upcoming US presidential election and other elections around the world are free and fair.
Facebook’s updated political ads policy
Facebook announced a series of updates to its ads policy, including changes to enhance transparency in its political ads library and to ban ‘deep fakes’. We welcome these incremental changes, which demonstrate Facebook is making some progress in meeting its obligations under the Contract for the Web.
But we remain concerned about Facebook’s failure to limit the micro-targeting of political ads. The company plans only minor updates, including giving users the option to choose to see “fewer” political and social issue ads on Facebook and Instagram, and to see no ads from campaigns that already have their contact information — both on an opt-out basis. These controls won’t meaningfully prevent advertisers from narrowly targeting different groups of people with specific ads, and so do little to address the broader danger of targeted political advertising. This leaves Facebook lagging behind its peers, including Twitter, which banned all political ads, and Google, which now allows just basic demographic targeting for political ads.
Why Facebook must ban micro-targeted political ads
The Web Foundation is calling for Facebook to suspend micro-targeted political ads globally. Social media is an important way for candidates to reach voters, but their targeting should be broad. It’s reasonable that political campaigns are able to reach voters using basic demographic information like age, gender and geographic location — just as they can tailor their messages to different, broad audiences in the offline world. But granular targeting of voters undermines the political process and Facebook must urgently update its policies to safeguard upcoming elections.
To have constructive, honest political debate and fair electoral campaigns, a society needs to have a shared understanding of what candidates and parties stand for. In the past that happened through manifestos, TV ads, radio spots and press conferences that thousands or even millions of people saw.
Microtargeting of political ads changes this model, allowing candidates to show ads with different (often contradictory or even misleading) messages to different users. This individualisation of ad targeting reached new heights in the 2016 US presidential election, where the Trump campaign served 5.9 million ad variations in just six months.
When ads are micro-targeted in this way, voters can have wildly diverging understandings of who the candidates are and what they stand for. It prevents people from evaluating the full platform of any candidate and therefore limits their ability to make an informed decision at the ballot box. And it also means it is extremely difficult to challenge or fact-check claims made in ads that may only ever be seen by a small handful of people.
Defenders of microtargeting may argue that it is consistent with existing political norms, and that candidates have always tailored different messages to different groups — for example voters living in a particular area or members of a local religious community. This may be true. But the difference here is the sheer scale and speed of microtargeting that is possible in the age of social media, as well as the vastly more sophisticated and granular options for targeting. As in many areas, the scale of Facebook’s responsibility must match the heft of its influence.
What microtargeting looks like
A recent review we did of Facebook’s ads platform showed that a political candidate would be able to target a small audience of potential voters in a specific city who have “some high school education,” are unemployed, and are interested in gambling. The campaign could share one message with this group and a very different message with an audience of high-earning college grads, with a new job, who are interested in luxury goods. These audiences could be further narrowed with dozens or even hundreds of additional demographic, behavioural and interest categories.
We believe that Facebook should limit its ads targeting to, at a maximum, the broad demographic criteria of age, gender and location. Facebook has already limited ad targeting in other sensitive areas. Last year the company made changes to restrict targeting for housing, employment and credit ads in the US, with advertisers no longer allowed to target users by “age, gender, ZIP code, multicultural affinity, or any detailed options describing or appearing to relate to protected characteristics.” In the case of political ads, we believe Facebook should similarly develop restricted targeting criteria for political ads globally, though the targeting categories that we believe are acceptable for political ads are different to those in housing, employment and credit markets.
Users are in the dark
Today, people seeing political ads are none-the-wiser that they are targeted for very specific demographic and behavioural reasons. Restricting targeting to basic age, gender and location information would mirror Facebook’s decision to provide transparency only around very broad demographic targeting criteria in its ads transparency tools.
Currently, even if Facebook includes every ad shown on the platform in its ads library, it fails to provide transparency beyond showing high-level breakdowns of who was shown the ad by gender, age and broad location. And Facebook’s “Why Am I Seeing this Ad” feature for individual users provides only very high-level targeting information, often limited to vague phrases such as “you’re seeing this ad because you’re between the ages of 18 and 55 and live in the United States.” The significant gap between the targeting options Facebook gives advertisers and the transparency it gives users further underlines the need to restrict targeting.
Calls for regulation
While announcing changes to its ads policies, Facebook repeated a call for governments to step in and provide regulation for online political ads. Though we support the urgent need for industry-wide regulation, particularly around online political campaigning, in the absence of such regulation Facebook must adopt policies that support democracy and that are in line with human rights standards — including stopping micro-targeted political advertising on its platform.
At a minimum, Facebook should publish a human rights impact assessment to demonstrate why it believes its current approach to micro-targeted political ads is not harmful to individual human rights, including the rights to privacy, free expression and association, among others.
With US presidential campaigns gearing up for 2020, there’s no time to spare.
To receive a weekly news brief on the most important stories in tech, subscribe to The Web This Week.