Hamsini Sridharan, Program Director, MapLight

Illuminating Dark Ads




Right now, digital advertising is like a one-way mirror: advertisers—including political actors and commercial brands—can “see” a user down to the pair of shoes she just bought, her political preferences, her race, age, gender, and more. Meanwhile, she has little information and even less control about the ads she is shown, who is paying for her attention, or why. On platforms like Facebook and Google, ads can be run so cheaply and targeted to such narrow groups that only the advertiser knows who is seeing what—meaning there is virtually no way to hold them accountable for their messaging. These “dark ads” are ripe for deceptive practices, from political manipulation to predatory or discriminatory commercial tactics, including those that target vulnerable or marginalized communities. They compound the political problem of dark money, making it harder for users to understand the real interests behind a message.


A centralized website or app that 1) compiles a public database of advertising (especially sensitive material such as political ads) from all platforms, providing meaningful transparency into who is being shown what and when and 2) allows users to customize their advertising controls across all platforms at the individual level. To build this tool will require an industry-wide ad metadata standard.  

Who is this for?

  • Social media and internet users with average levels of tech fluency

  • Journalists and researchers

  • Government agencies (such as the FEC and the FTC)


The public advertising database will have two functions. First, it will allow individual users to search all ads targeting them. Second, it will give society’s watchdogs, including journalists and regulatory agencies, greater transparency into what types of targeting are being conducted in order to identify how to protect the public from harm. The database will capture ads across digital platforms, with a focus on critical areas such as political advertising. To do this, we would establish an ad metadata standard. Each database entry will include data on who is running an ad (pinning it to a real entity rather than a shell name), how much it cost, what data sources were used, when and where it was shown, and, critically, the specific audiences and keywords targeted. Data would be available for systemic analysis via an API.

We simultaneously need to make it easier for ordinary internet users to have full, opt-in control over how they want to be reached. Instead of advertisers creating audiences (21-34 year-olds, people interested in progressive politics, dog lovers, etc.), users themselves should be able to choose which audiences they want to join. Say that a hypothetical user, Malini, uses Facebook, Twitter, and Google. Right now, these services determine Malini’s interests for advertisers (and only allow her to opt-out via obscure settings). With this tool, she decides what categories of advertising she wants to see and what criteria advertisers can use to reach her. Most importantly, she can do this all in one place, rather than having to go through the settings of each platform.

Related Resources