Q&A on Transparency for Electoral and Issue Ads

Our new ads transparency tool for electoral and issue ads launched in the US today. Advertisers placing political ads are now required to verify their identity and location. Once authorized, their ads will appear with a label in News Feed, disclosing who paid for the ad from the advertiser, and be available in a searchable archive for up to seven years. Anyone on Facebook can now access the archive too. Below is the transcript of the press briefing call we hosted with reporters to detail the new developments.

Stephen Satterfield: Hi, everybody. This is Steve Satterfield and I’m a Director on Facebook’s policy team.

Last fall, as you know, Mark announced a number of steps that we’re taking to protect elections from bad actors on our platform. In March, we gave a broader overview of our efforts so far and promised to provide regular updates on our progress. Today we’re going to talk in detail about one of those updates and then we’ll spend the rest of the time answering your questions.

As a place to start, though, I want to emphasize how seriously we’re taking this issue and the resources we’re dedicating to it. We’ve got a number of new initiatives designed to create a robust approach to protecting election integrity on Facebook and Instagram.

And they’re driven in part by what we’re learned from recent elections in places like Germany, France, Kenya, the United Kingdom and also through collaboration with partners like the Atlantic Council and the Comparative Agendas Project.

Now when it comes to this work, there are five core areas where we’re focused. Those are:

  • Combating foreign interference,
  • Fighting fake accounts,
  • Boosting investments and security,
  • Reducing the spread of false news, and
  • Increasing ads transparency — and that’s where we’re going to share more detail today.

Simply put, we think ads should be transparent; it’s one of the core ads principles that we shared last fall in a post by our Vice President for Ads, Rob Goldman. You should be able to understand who’s showing you ads and you should see what other ads that advertiser is running.

Now we heard feedback from people that they wanted to see all the ads that were running from a Page; not just those that were targeted to them. We responded by building View Ads; that’s a product that’s testing in Canada and Ireland and that we’re going to rollout to everybody next month. And again, this is something that you can go to an advertiser’s page and click to see all of the ads that that advertiser is currently running. Today we’ve got news on more ads transparency efforts. And these are launching today.

I’m going to turn it over to Rob Leathern, who leads the work on these transparency efforts for our ads team to give you the news.

Rob Leathern: Thanks, Steve. Hi, I’m Rob Leathern, Director of Products on the ads team.

Starting today, all election and issue ads on Facebook and Instagram in the US must be clearly labeled, including a “Paid For By” disclosure from the advertiser at the top of the ad. This will help ensure that you can see who’s paying for the ad, which is especially important when the page name doesn’t match the name of the company or person funding the ad. This also meets the commitments we made back in October to increase the transparency of the election related ads you see on Facebook.

When people see that label, it means the person writing the ad went through the authorization process and verified his or her identity and location. We believe this new level of transparency is good for people and will allow journalists, researchers, NGOs and others to hold campaigns, candidates and organizations accountable for the ads they create.

And all people on Facebook, no matter where they live, will also be able to access and review a searchable archive that will house these ads for seven years from the day they run. People that visit the archive can learn things like range of spend and impressions delivered and demographics reached like age, location and gender for any political or issue ad. The archive is searchable and you can filter by page name or active or inactive status. We’ll also expand the authorization process in the ad archives to other countries in the coming months.

We think these new features set a new standard for transparency in digital advertising and they are a key part of our broader elections integrity work that Steve mentioned. We believe the process we put in place is a solid step, but we also know that initially there will be instances where we don’t catch ads that should have been labeled and the authorization process wasn’t completed by the person placing the ad.

That’s why we’re asking for people’s help. If a person sees a political ad that doesn’t have a label and needs one, we’re asking them to please report it. The way people can report it is by tapping the three dots at the top right of the ad, selecting “report” and then selecting, “it refers to a political candidate or issue.” We’ll review the ad and if we determine it has election related or issue content, we’ll take it down and add it to the archives.

This is important because now even ads without a label that are deemed to have political content will be housed in the archive. The advertiser will also be prevented from running election-related and issue ads until they complete the authorization process. And we’ll follow up with you on the outcome of the ad that you reported. This is a tool that makes it easier for you to find problems and that’s something that we want. We invite you to report ads so we get better, faster.

We also know that outside experts, researchers and academics can help by analyzing political advertising on Facebook, and that’s why we’re working closely with our newly formed Election Research Commission and other stakeholders to launch an API for the archive.

Another area where we’re working is on the news coverage of elections and important issues. We recognize those are distinct from advocacy or electoral ads, even if those new stories receive pay distribution on Facebook. We’re also announcing that we’re committed to updating the archive to help differentiate between news and non-news content.

Now let me turn it over to Katie Harbath on our policy team to walk through how we’re thinking about defining ads with political content on the platform.

Katie Harbath: Thanks, Rob. Hi, I’m Katie Harbath and I help lead our global government and elections effort.

There’s two pieces to today’s launch that I want to talk about. First is the product, you just heard Rob talk about our new ad transparency tools and requirements; second are the new policies we’ve created to educate people and advertisers on this new process.

When you want to give people more information about political content, the obvious question is what do we consider ‘political?’

The answer might seem to be that ads that advocate for a certain candidate or cause, but just looking at ads on either side of a candidate race would not go far enough.

And after making that initial announcement in October, we heard feedback from a range of independent stakeholders and third parties that we should also include issue ads or ads about important topics being discussed across the country.

After all, many of the Russian ads around the election focused on stirring passion around divisive social issues without mentioning any race or candidate. To shape these policies, we first look to current and proposed laws around ad transparency. This was helpful guidance but these laws define issues more broadly, such as, quote, any issue that is of national importance.

In order to enforce on a policy, we had to take a different approach and be more specific about defining what those issues are. We started with the nonpartisan Comparative Agendas Project to help build our foundation. For decades, this group has worked to assemble and code information on a policy processes of governments from around the world.

We also reached out to many other advocacy groups, and in working closely with these groups, we developed a list of 20 initial issues — like, health, public infrastructure, civil rights — and have crafted a policy. We know this needs to be a living and breathing document, so we’ll continue to seek feedback on new topics and issues working with YouGov, a group that continually collections opinions from across the world regularly.

The aim is for this effort to reflect the latest public discourse as we continue to update it. To enforce the policy and affect which advertisers are running ads with political content, the ad review process looks at the text and image of an ad, for example. We also use machine learning to help identify these ads and prevent them from running if that advertiser hasn’t been authorized.

And we are planning to hire 3,000 to 4,000 more people to review ads and our machine learning improves the more ads we review. As Rob said, we won’t always get it right. We’ll likely miss ads that aim to persuade and may label others when they don’t have political content; we’ll keep working on the process and improve as we go.

There’s a lot to learn, so we wanted to share a few examples to illustrate how we’re thinking about issue ads that will require authorizations and a label. For instance, on the topic of education, ads regarding points of view on student loan policies would require a label, but advertisements on a particular university for student enrollment would not.

On the issue of taxes, ads taking a position on whether taxes should be raised would require a label, while an ad about a wealth management app would not.

On the topic of crime, ads supporting legislation to limit prison sentencing would require a label, but an ad promoting the season premiere of a real-crime TV show would not.

Deciding what is or is not an issue is inherently controversial and not everyone will agree with our approach, but our intent is to help people understand who is trying to influence them when it comes to political and social issues and why, and this is a significant step for us in that direction. In the coming months, we’ll expand our ads transparency features to other countries and work with international partners to determine what issues apply.

And now, I’ll ask Steve wrap it up.

Stephen Satterfield: Thanks, Katie. To sum up here, the transparency these tools will bring to ads with political content is a key part of our approach to election integrity and it’s also something that we’re going to continue to work to improve. We hope this was a useful discussion and we’ll plan to do more of these updates as we continue to make progress on our efforts.

And now, we’d love to open it up for your questions.

Operator: Your first question today comes from the line of Jonathan Vanian from Fortune. Your line is open.

Jonathan Vanian: Hey, guys. Thanks for having this. The 3,000 to 4,000 more people, is that — is that additionally to the 20,000 security staff you were planning on hiring by the end of the year? And when do you expect the 3,000 to 4,000 people to come onboard?

Rob Leathern: That 3,000 to 4,000 number is included as part of the 20,000. We’re not breaking out timing specifically on when we expect those folks to be on board but it’s an ongoing — an ongoing process of adding them between now and the end of the year.

Operator: Your next question comes from Tony Romm from the Washington Post. Your line is open.

Tony Romm: Hey. Thanks so much for doing this. Just two questions for you. First, I know this is just the beginning of things you guys plan to roll out, but are you looking in the political ad transparency hub to make it possible for folks to see aggregate numbers? I could see a member of Congress and the total number of individuals they reached and the total number of like particular categories of individuals?

And second, do you guys plan to create a new feature down the line where you make it clear what interest categories advertisers sought to target their ads to? Right now it just shows each demographic of person who saw the ad, but not maybe the audiences to which they sought to target that ad. Do you guys plan to address those two things in the future? Thanks.

Rob Leathern: More generally, we’re obviously eager to hear people’s feedback using these tools and see what is useful to them. Currently, we’re showing you the information, the demographic breakdown of the people who actually saw the ad, which we believe gives a really good indication of the actual impact of the ad that has political content and offers more transparency than the intent in terms of showing the targeting.

We really think demographics like age, location and gender are a great first step. We’re going to continue to explore and evolve the archive as we learn more about how people use it. We also know that advertisers test lots of different versions of ads. We’re also considering as we continue to grow and work on this product ways that we might give additional views into this data.

And then, the final point I’ll bring here is we really think providing an API to researchers and others will allow them to conduct their own analyses on this data. And so we think that’s a really important step as well.

Operator: You’re next question comes from Josh Constine from TechCrunch. Your line is open.

Josh Constine: Thank you. Do you plan to work with any third party auditing services to ensure there’s no bias in the definition of what constitutes political content in advertising?

Katie Harbath: Hey, it’s Katie. I think that’s partially why we’ve been working with the Comparative Agendas Project and also talking to many different groups across the political spectrum about the definitions that we’re working on. And we will continue to do so.

Operator: Your next question comes from Steven Levy from Wired. Your line is open.

Steven Levy: Yes. Thanks for doing this, folks. Two questions. One is sometimes in the political realm organizations pop up, PACs with names that are kind of misleading that really wouldn’t provide much context, even when people knew who was paying for the ad. Are you just going to just take those and some of these PACs that might have come from dark money?

And second, are you making special provisions for people who might be reporting ads which are legitimate, but just people on the other side of those ads who might want to get you involved in questioning or taking down the other ads?

Rob Leathern: I can address the second part of the question, which is about people reporting ads. Obviously we recognize that not everyone who reports an ad would necessarily have the same intentions of improving the integrity of the process. That is why we’re not just taking ads down or marking them as political if they’re reported. They’re also subject to review both by our machine learning models as well as by people — our reviewers.

It’s very important that we are not just taking people’s word for it, but we do believe that the input that people have on ads — either reporting them if they’re in the archive or reporting them if they’re not in the archive — it’s extremely important. We’ll continue to grow and evolve our enforcement and work around that area.

Stephen Satterfield: I would also add this is why we’re developing objective standards for evaluating whether an ad has political content in it — whether it meets the definition that we are — that we have in our policy. It shouldn’t matter, the intent behind why somebody is reporting; if the ad meets the criteria under which we would consider it an ad with political content, it must bear the label and be archived.

Katie Harbath: And let me answer your first question. That’s exactly why we are having the archive — so that we can help shine more light on these ads and so that watchdog groups, reporters, election regulators are able to see these ads and then know more about who’s actually behind them.

Operator: Your next question comes from Kurt Wagner from Recode.Your line is open.

Kurt Wagner: Hey. Thank you very much. Can you give us a better sense of how you’re going to balance AI versus human review with this process? If I’m a registered advertiser, are humans mainly reviewing all the ads I give you? And then what role is AI is going to be used in finding people that are trying to get around the system that you guys have in place here?

Rob Leathern: This is Rob. Obviously machine learning and AI is extremely important, but so is human review, which is why we’re adding those 3,000 to 4,000 people that we mentioned specifically in this area. And they have to work — they would have to work together. One does not — is not going to be sufficient without the other. We’re going to continue to look at various signals.

We do incorporate the signals we get from people reporting ads as well. We will look at other aspects, other data to really understand how to find political ads. As Katie mentioned, we’re also working with third-parties with respect to how we define the issues; that’s obviously going to be subject to ongoing feedback.

Some of these as you can imagine will be difficult to nail down. There will be mistakes that we will make in the process, which is why we’re going to hold ourselves accountable. We’re providing an API, we’re going to bring others into the fold — third parties to help work on helping us understand how well we’re doing with this.

Operator: Your next question comes from Casey Newton from the Verge. Your line is open.

Casey Newton: Hey, I wanted to ask what’s going on with publishers. I know you said you’re working with them. My reading of your blog post is that you are going to require them to put Paid For labels if they boost a post, but then there was also something about setting up a separate archive for those posts. Can you just walk us through your thinking on that and why you think it’s important that publishers label their advertisements?

Stephen Satterfield: This is Steve. I can take that question. Thanks. Any ad that has political content on Facebook going forward will require authorization, labeling and archiving — regardless of who’s running it. We think it’s important that any ad that mentions a candidate for example include this transparency and be accessible via the archive. We have heard feedback, particularly since we announced the expansion of our policy in April from various groups who were concerned about being required to be labeled and maintained in the archive.

And so we are in ongoing dialogue with those groups, including with news publishers. We’ll have more to say about that in the future. We appreciate the feedback and we are taking it into account as we move forward.

Operator: Your next question comes from Ian Sherr from CNET News. Your line is open.

Ian Sherr: Hi. Thanks for doing this and having me on. I’m piggybacking on what Kurt was asking, about how well is the AI working at this point? Or are you even using it? And I’m asking because we heard mixed results about stuff like identifying hate speech in other ways and so I’m curious how well-trained the AI is at identifying the political ads that then need to be checked up on.

Rob Leathern: Yes. This is Rob. I can answer on that one. When it comes to the machine learning, obviously issue ads are newer to look at than the electoral ads. If you recall, our original announcements were around electoral ads. We’ve obviously been looking at those very closely for some time. We believe that we have very good coverage in place for all of these. But as I mentioned, we’re going to have to work with outside parties and researchers to continue to grow and evolve our understanding of these kinds of ads.

I’d say that part of how we’re going to be held accountable and how we’re going to involve folks is by providing the archive, which will show you all of the different ads that we have — we are seeing and interpreting as political issue content. I think that’s really important, is that the archive shows how we are doing and we’re obviously looking for feedback from people. And then secondly the ability to report an ad — any ad on Facebook as of today can be reported as a political ad and then we will have people and models review those ads to determine whether they are, in fact, political in nature.

Operator: Your next question comes from Chiara Sottile from NBC News. Your line is open. Chiara, your line is open.

Chiara Sottile: My question’s been answered. Thank you.

Operator: Your next question then will come from Anna Hensel from VentureBeat. Your line is open.

Anna Hensel: Hi, thanks for doing this, guys. My question is how you see research functions evolving over time? Immediately I thought of a voter, for example, who might want to see what candidates in their area are running ads but might not immediately know who those candidates are.

Rob Leathern: This is Rob. We have a number of other product initiatives around providing information to people about the candidates in their area and running in their local elections. Over time, I think it’d be fair to see us work to find the right mix of bringing these different products together. We’re definitely looking for feedback and interested in hearing the response of people using these different products to see what makes the most sense.

One other thing to point out is that you’ll be able to get to the archive from ads that are labeled in your News Feed. There’s one example of where you’re going to go from the individual ad, if you are in the audience that is going to see that specific ad, and from there you can get to the archive, see other ads that are from the same advertiser, same political advertiser. And from there you can also start to explore and see and search for ads from other candidates.

Katie Harbath: To piggyback off that, I would say too, those products that Rob mentioned, we have one now, Town Hall. And in 2016, we also did one where we showed people who was on their ballot. And so we could easily see people using that to go see who represents them, and then going to the archive and searching for their name.

Operator: You next question comes from Tim Johnson from McClatchy. Your line is open.

Tim Johnson: Hi, I listened with great interest when Mark Zuckerberg talked about the number of fake Facebook accounts that have been attempted to get set up — 580 million; that’s 6 million a day. Do you expect a tidal wave of questionable ads just as you’re dealing with fake account set ups? Are you already facing that? And what are some discrete numbers in that regard?

Rob Leathern: We’re not sharing numbers about numbers of ads specifically. But we have a number of efforts running that are — obviously, some related to political ads, some not — that’s addressed these. Obviously stamping out fake accounts is a very important piece of work here. Mark has shared previously as well that we’re involved in various different games of cat and mouse with bad advertisers of all kinds.

We have a lot of folks focused on this in addition to the 3,000 to 4,000 additional reviewers that were mentioned earlier. We also have up to 20,000 folks this year that are working in these areas. We do believe that we have a lot of very strong programs in place to address these. And we also look forward to sharing more information with the public about our progress in this area.

We don’t have anything to share specifically on the ads part of it right now though.

Katie Harbath: I would add that any fake accounts or anything wouldn’t even be able to create any of these political or issue ads because they would have to go through authorizations first or they would have to prove that they are who they say they are as well as get that code in the mail and put that in. And so that’s part of the whole reason that we put in this authorization step.

Operator: Your next question comes from Max Cherney from MarketWatch. Your line is open.

Max Cherney: Hi. I’m wondering why, will, or won’t — and to what extent — the changes to ads transparency will have on ad revenue as well as the impact on engagement and other metrics your advertising clients watch closely?

Katie Harbath: I think, at Facebook, our work around civic engagements and in elections encompasses a lot of different things including our election integrity efforts as well as our civic engagement products. And our ultimate goal is to help make sure that people know who’s trying to share messages with them, as well as who is paying for them.

And Mark has said, we won’t make money off of political ads And in fact, we’re investing a lot into this. And the cost is worth it for us on that. And so we are building this for a lot of different reasons. And ultimately, it’s to make sure that people have a way to express themselves and engage on political discourse on our platform in a transparent way.

Operator: Your next question comes from Ali Breland from The Hill. Your line is open.

Ali Breland: Hi. My question was about, if these sort of changes aren’t successful would you guys or have you already considered the option of doing preemptive human review specifically for political and issue-based ads given that they do seem to be more important and you guys seem to think they’re more important as well?

Rob Leathern: When you say preemptive review, I’m not 100% what that means. However we are committed to — when we look at advertisers in this area, obviously people are self-identifying as political advertisers. If someone is running an ad that has political ad content and we flag it either via machine learning models, either via people pointing it out to us, and then us confirming it, or by the ongoing review that we do of ads, we will then add it to the archives. And we will make sure that not only is that version captured in the archives, but we will then put the advertisers through the authorization process.

If they are unable to go through an authorization process, that ad will still remain in the archive so that people can see it. If that ad also violates other ad policies of ours, we’ll actually — it will still be in the archive and people will be able to see that as well. Once that entity is able to successfully authorize, presuming that they are, they could then start running the ad again and it would — then that second version would also be in the archives.

We want to just make it extremely clear to people what is being run. We’re not only capturing ads where the advertiser has self identified as being a political advertiser, but also ones that we flag that are running that have not been necessarily a priori designated by the person’s political ads. We think this is really good for people and we’ve also gotten a lot of feedback from good actors that they believe that these steps are valuable and will help to protect their own work as well as to protect elections.

Operator: Your next question comes from Janko Roettgers from Variety. Your line is open.

Janko Roettgers: Hey, guys. I’m wondering if you could give me a little bit more details on the archive on day one as it is today. Are you basically starting from scratch or are you refilling it from its past — as people had in the past couple of months or even years, and are you going to add Internet Research Agency ads to that as well?

Rob Leathern: I’ll take the first part of this. This is Rob. We started adding ads to the archive in early May, starting with ads from advertisers who completed the authorization process. Any ads that started running — they were running before that time wouldn’t be reflected in the archive. We also are now adding ones where we have determined that it’s political where someone may not have self-selected.

We’re also aware that it takes — because of the authorization process — it may take a few days or a week to get through that process and so we are giving people a short window of time to make sure they go through this process before we start — before we prevent the advertisers from running those ads. This will just only be a very short period for that — about the next week or so — before we then will prevent all advertisers that we flag as running political ads from being able to run without the authorization.

Stephen Satterfield: And on the IRA ads question, as you know we provided those ads to Congress to the congressional committees who are running the investigations into potential interference with the 2016 election. And those ads have been released to the public. We are focused on efforts going forward to preserve election integrity.

Operator: Your next question comes from Ishmael Daro from BuzzFeed News. Your line is open.

Ishmael Daro: Hey, guys. Thanks for doing this. I had a question about the API that you’re making available for researchers. I’m wondering just how granular that data will be compared to the public-facing archive? And then also how you are going to regulate access to the API in light of the Cambridge Analytica stuff? Thanks.

Stephen Satterfield: Thanks for your question. We’re figuring these things out right now in connection or in conjunction with a number of partners who are very interested in getting API access. We’re really excited about the prospect of building this tool because it will give watchdog groups, researchers, academics and others the kind of robust access to the archive that they’ve been asking us for. Once we have more details to share about the archive, we’ll share those with you.

Operator: Your next question comes from Thomas Claburn from The Register. Your line is open.

Thomas Claburn: Yes, hi. Thanks for taking my question. The product ads have content requirements. Are you going to be doing any kind of fact-checking or imposing any kind of content requirements on political issue ads or can people say pretty much whatever they want?

Katie Harbath: We have our content standards that all ads — including these ads — are going to have to follow in order to run on the platform.

Stephen Satterfield: Those policies include a requirement that ads not be deceptive or misleading. And if we became aware that there was deceptive or misleading content in an ad, we would take action against it.

By the way, enforcing this policy — transparency is a core way in which we will improve in enforcing this policy. Not only will we potentially review an ad and the content of the ad; we’ll ask others or we’ll make available the ad to others who can also report the ad and flag things that potentially could violate our policies.

Rob Leathern: This is Rob. One other thing I’d add as well is that next month, we’ll be launching View Ads transparency, which, in addition to the political ad archives, actually extends transparency to all ads on the platform globally. From there, people will also be able to report ads that are misleading or deceptive. And so we believe that this step as well will increase the ability for people to make us aware of ads that they believe are not in good faith.

Operator: Your next question comes from Edgar Alvarez from Engadget. Your line is open.

Edgar Alvarez: Hi. Is there a plan for the searchable archive to include political ads from elections outside of the US? And then the second part to that is and what was the thinking behind the up to seven years timing? Thank you.

Katie Harbath: I can take both those questions. We are planning on rolling these tools out more globally in the coming months and so over time this archive will include ads from other countries. And the thinking around the seven years is we wanted it to include a full Senate cycle, which is six years and also in other countries, the longest period between elections could be potentially six years as well.

Operator: Your next question comes from James Hercher from AdExchanger. Your line is open.

James Hercher: Hey. Appreciate you guys taking the question. Mine’s on fundraising, which is a big focus for political advertisers when it comes to Facebook. I just wonder if that’s something that you’re approaching as its own topic, in a sense, when you’re looking at these political ads transparency initiatives, if you’re looking at fundraising calls or people who are linking out from Facebook?

Katie Harbath: Fundraising ads by a candidate, political party, political action committees or issue-based organizations very likely will be requiring their label. It’s hard to speak hypothetically of every type of instance but I think you can very much expect to see fundraising ads in the archive.

Operator: Your next question comes from Eric Lieberman from The Daily Caller. Your line is open.

Eric Lieberman: Hello. I was wondering how long these reforms have been in the works and if they were in relation, in any way, to recent proposed rule-makings by the Federal Election Commission to include websites and internet-based applications into their rules against political disclaimers?

Katie Harbath: This is Katie. We’ve actually been working on this for quite some time. We made the first announcement in October that we would be doing this work and then we updated it in April that we were going to be including issue ads. And we’re taking these steps regardless of where legislation may go and we’re continuing to work with partners as they look at that.

Stephen Satterfield: Yes, we’re certainly listening to developments that are happening on the Hill, including proposed legislation like the Honest Ads Act and we’re very aware of the proposed rulemaking of the FEC. We’re taking all of these things into account. You can see that our policy reflects language from existing laws as well as proposed laws.

Katie Harbath: And I would also just say too, many campaigns have been adding disclaimers to their ads already in past elections as well. And so we’re also — this helps make it easier for them to be doing so and for people to be able to see those disclaimers of who is paying for the ad.

Stephen Satterfield: Yes. And I think another thing to note is that yes, there are a number of developments that are going on at the FEC and on the Hill — that is, in Congress — but we’re not waiting. We’ve been hearing calls for increased transparency around ads with political content for some time now. We’re excited that today is the day when we’re taking the first step toward providing the transparency. And we hope others follow.

Operator: Your next question comes from Jeremy Merrill from ProPublica. Your line is open.

Jeremy Merrill: With regard to the human beings who are going to be reviewing the ads that are flagged by the machine learning system or that are reported by Facebook users, will you all be sharing the objective standards and the instructions and the training materials that these human reviewers are going to be receiving?

Rob Leathern: This is Rob. I think it’s fair to say we have obviously chosen to do that on the Community Standards side recently. Certainly something we would consider over time. I think one of the concerns in general we have on the ads side more generally speaking here, just kind of up-leveling it beyond just political ads, is that we face a lot of bad actors who try to game our enforcement systems.

Obviously anything that we would share here we have to just be cognizant it may be used to try to either get up to a line that is a gray area or to really, like attack us. And we are giving — as part of the ad process — we’re giving people feedback immediately about whether their ads are getting through or not. We’re definitely in a cat and mouse game with bad actors, so we just have to keep that in mind. But certainly it’s something we would consider over time.

Tom Reynolds: Hey, Operator, this is — oh, I’m sorry, Steve. I didn’t mean to cut you off, go ahead.

Stephen Satterfield: No, I just wanted to add that we do recognize that we need to be clear with people and with advertisers about the criteria that we’re using to determine whether an ad is an ad with political content. That’s why we’ve developed a list of 20 issues that essentially elaborates on the higher level policy that we’ve developed. I think we’re going to continue to refine that list over time.

And we’ll continue to provide guidance to advertisers — information to people — about how we are making these calls. We’ve tried to define them objectively, the kind of ads that will be covered. And you’ll of course be able to see the results of the review and advertisers’ self-categorization in the archive. And we, of course, would welcome feedback on that as well — the kinds of ads that you’re seeing.

Tom Reynolds: Operator, we’re going to have time for two more questions, OK?

Operator: OK. Your next question comes from Sheera Frenkel from New York Times. Your line is open.

Sheera Frenkel: Hi, I actually had two questions. One is if this is going to be broadened out to apply to Instagram as well? And the second is, when you talk about applying this globally to other elections, what your verification process going to look like? Obviously, there are countries where mailing something to someone’s home or office to confirm their identity isn’t going to work. Are you working with civil (disobedience) groups to come up with a policy in the rest of the world?

Katie Harbath: Hey, it’s Katie. On the first one, this archive does include Instagram ads. This does apply to Instagram today. And so you’re going to be able to see those Facebook and Instagram ads in the archive. And the label is also applying to both.

Globally, we’re going to be working with election regulators and organizations in different countries on everything from the authorizations process and also how to be defining what an issue ad is. And so that’s part of — it takes time to be rolling these out globally because we want to make sure we take the time to do that — and most importantly, to get it right — for what the system is in that country.

Rob Leathern: One other thing I’d add to that as well, is that, we also, in other countries as well, will be rolling out the View Ads globally, View Ads transparency starting next month. In this area also, we’re going to emphasize that we want people in various countries to reporting ads that that they believe are political and we’ll use that as a way to also expand more quickly into other countries that we may not have the political ad archive and authorizations process covering yet.

Operator: And your final question for today will come from Fergal Gallagher from ABC News. Your line is open.

Fergal Gallagher: Hi. I’m just wondering how easily searchable the archive will be. Will people be able to search to see which ads have reached the most people in their location or by gender, et cetera, or how will people find ads?

Rob Leathern: Thanks for the question. This is Rob. Today, the first version of the archive, people will be able to search and filter by keyword or Page name and they can also then see whether the ad is active or inactive. And over time, we’re going to get the feedback from people who are using the product, conduct more research and help to understand what kind of additional features we should add.

Then in addition, the API will really be one of the ways in which we really get external researchers to provide their kind of way looking at things whether that’s drilling down on demographics or location, et cetera. We do believe that when we release that API that that’s also going to be a great additional way for people to get more insights through those analyses.

Tom Reynolds: All right. Operator, I think with that, we’re going to wrap up. Let me just say thanks, everybody.


Audio File of Press Call 

To help personalize content, tailor and measure ads, and provide a safer experience, we use cookies. By clicking or navigating the site, you agree to allow our collection of information on and off Facebook through cookies. Learn more, including about available controls: Cookie Policy