LONDON — Facebook faced a wave of criticism for missing interference by Russian-based groups during the 2016 presidential campaign in the United States. Now, it is making Ireland an early test case for new policies meant to block such meddling in foreign elections.
With a contentious May 25 referendum on Ireland’s abortion ban approaching, Facebook said on Tuesday that it would block political advertising from groups based outside the country. The company also recently introduced a tool so users can see all the ads a group is posting on the social network, in a bid to increase transparency of political campaigning on its platform.
Facebook has said similar tools will eventually be rolled out in other countries, with analysts and observers focused in particular on the 2018 midterm elections in the United States.
The company’s response illustrates its more aggressive approach to regulating political advertising since the 2016 campaign, when, the company disclosed, it sold more than $100,000 worth of political ads to Russian-linked accounts.
“Our company approach is to build tools to increase transparency around political advertising so that people know who is paying for the ads they are seeing, and to ensure any organization running a political ad is located in that country,” Facebook said in a blog post outlining the Irish decision.
The referendum on whether to end Ireland’s constitutional ban on abortion, a vote that is expected to be tight, has drawn concerns that foreign groups will try to influence the debate through social media. Irish electoral law bans political donations from non-Irish citizens or residents, but online advertising was not envisioned when the law was written, and transparency activists and online monitors have reported a swelling number of online ads, many of unknown origin.
Facebook said on Tuesday that groups working on both sides of the Irish referendum would be able to flag advertisements suspected of coming from foreign organizations. The advertisements will then be investigated by the social media platform. The company also said it would use artificial intelligence technology to spot potentially problematic material.
Many countries, including the United States, prohibit foreign groups from advertising in domestic elections, but regulating the spending is difficult with more political activity moving online. Facebook’s advertising system has become a favorite of political groups because it is largely automated, and makes it easy to target narrow segments of voters.
Since the 2016 election, Facebook has made several policy changes to address concerns over the role it plays in elections and politics around the world. The company has changed its News Feed algorithm to de-emphasize political news, and has hired thousands of moderators globally to spot rumors and extremist content.
The company is also changing its advertising policies, including allowing only authorized groups to buy political ads. A new verification process will eventually require advertisers to be based in the country where the election is taking place.
Newsletter Sign Up
In addition, Facebook is planning to introduce a searchable database to show how much an advertiser is spending, as well as the demographic details of the audience that a group is trying to reach.
Padraic Ryan, projects coordinator for the social media verification service Storyful, said that Facebook’s move was a welcome one. But he added that countries like Ireland needed to formulate their own rules, rather than rely on private companies.
“This is about regulation and transparency,” Mr. Ryan said. “It’s not just about Facebook.”
He added: “We’ve seen preroll ads on YouTube, display ads on any number of sites, suggested results on Google linking to pro- or anti- pages, and it’s not at all clear who is behind these. It seems manifest now that there should have been rules regulated before the fact.”
Google and its video-sharing service, YouTube, did not immediately respond to a request for comment.
Europe in particular has become a testing ground for some of Facebook’s changes to protect against political interference. Before the German election in September, for example, the company deleted thousands of fake accounts and worked with election officials to stop the spread of misinformation more quickly.
Researchers are now focused on the referendum in Ireland, and the platform’s role in the campaign.
Last week, Gavin Sheridan, a former employee for Storyful, was able to trace one web page, ostensibly an information source for undecided voters but with no verifiable identification or contact details, to conservative Roman Catholic groups in the United States. The page seemed designed, Mr. Sheridan said, to draw in undecided voters, who would then be targeted with personalized ads for the campaign that seeks to maintain Ireland’s conservative abortion laws.
“It’s not necessarily underhanded to try and identify targets for advertising, but if you are not being transparent about who you are representing, then it’s a problem,” he said.
Liz Carolan of the Transparent Referendum Initiative, a group set up to monitor online content during the campaign, said that Facebook’s decision to ban foreign-sourced advertising was a step in the right direction, but that more needed to be done.
Facebook’s new Ad View tool, which allows people to see all the ads that a Facebook page is running — and not merely those targeted at them — is an improvement, she said. But it allows users to see only current ads, so it is impossible to gauge the totality of an organization’s online activity. Nor is it clear what proportion of the total amount spent on Facebook ads is made up of foreign-sourced ads of the type now banned by the social media company.