United States vs Europe on Tech
A tale of two different approaches to regulation
I’m later than usual in getting out my Wednesday newsletter as I wanted to watch the Energy and Commerce hearing today on Section 230. As I write this we only got through the first of two panels and they’ve just started the second. And frankly, I don’t think much new came out of either of them.
This morning while I watched the opening remarks I started to comb through the European Commission’s proposed rules around political advertising that they released last week. I continue to be struck at how different the United States and Europe are pursuing tech regulation, and how much what one does will affect the other.
If you don’t want to read the full 41 pages of the legislation Mark Scott and Clothilde Goujard of Politico have this great primer. Casey Newton yesterday also covered how European law is affecting the US when companies like Twitter decide to enact changes meant to comply with EU law to the entire globe.
This is a phenomenon for people in the U.S. to pay more attention to. Europe is way ahead of the United States in having any actual momentum on their regulations. It’s highly likely that something will be passed next year well ahead of the 2024 European Parliament elections. These changes could very well impact how platforms manage elections elsewhere around the world.
One area that Europe is ahead of is the regulation of political advertising. While policymakers in the United States are focused on Section 230, antitrust and child safety, the Honest Ads Act - one of the first pieces of legislation introduced to regulate tech - has gone absolutely nowhere.
In some ways, this makes no sense as I would think that regulation of political ads is something that we could get more agreement on across party lines. But alas we are not there.
However, this gets us back to how decisions by Europe might actually end up defining the rules here because platforms won’t want to have different products for different regions. To that end, I wanted to share some of my top-level thoughts on the proposed rules from the EU.
First, a little history and context as to why we need regulation for political and issue ads online. When online advertising first began being used by campaigns in the 2006 - 2008 time frame the regulations and rules that governed ads on television, radio, and print weren’t updated to include online. The FEC issued a few advisory opinions but nothing was ever put into law.
Fast forward to 2017/2018 when questions arose about how President Trump used platforms like Facebook to win, revelations of how the Russian Internet Research Agency used ads to manipulate voters, and the Cambridge Analytica scandal the platforms decided to build their own measures to prevent foreign purchasing of political and issue ads as well as provide more transparency overall of those political and issue ads running on the platform.
You might wonder why I keep writing political AND issue ads. There’s a reason. Traditional definitions of a political ad are pretty narrow to ones specifically in support of a candidate or encouraging people to vote. However, most of the ads the Russian’s ran were about issues such as immigration and Black Lives Matter. Thus, we needed a policy to make sure we were capturing those. More than just that, there is a lot of disagreement over what is considered “political” today. Sara Fischer in Axios covered this well in this post from 2019.
Moreover, we had to flip the concept of campaign finance and who had to comply with these rules. Traditionally, the rules would start with who the entity was. If you are a candidate, political party, Super PAC, etc then your ads must follow these rules. However, bad actors aren’t going to self declare that they are running these ads. So we had to start with the advertisement, look at if the content of said advertisement fits the definition of a political or issue ad, and then require the entity to comply. That leads to a lot of organizations such as non-profits or corporations who do not consider themselves political being included.
On top of all this at Facebook, we also required people to be located in the country where they wanted to run these ads. This became a big problem for Europe in the 2019 elections when the Commission, Parliament, and EU parties wanted to run ads across all the EU countries, but couldn’t. Facebook ended up carving out some exceptions, but this issue is part of the reason the Commission says they wanted to update the rules.
This brings us to the proposed rules. Overall, these are really comprehensive and I can tell they did consult a wide range of perspectives in writing it. A few things I would point out:
Very limited scope: While this is somewhat comprehensive, the Commission points out that the recommendations do not address the content of ads including what to do about falsehoods in ads or blackout periods.
Research/Assessments: In these proposals, the Commission would require companies “to carry out publicly available assessments on how political ad targeting would impact elections in specific countries, as well as outline how political groups use their networks to talk directly with voters.” I haven’t seen this requirement anywhere else and would very much welcome it IF outside researchers were employed as well to give these assessments. We shouldn’t just rely on the platforms to give us the full picture as they will want to make it look as positive as possible.
Defining what ads are in scope: The Commission is asking platforms to look at a variety of things to determine if an ad will influence the outcome of an election including the context of the message. This is extremely hard to do at scale right now. AI is getting better at that - something Facebook wrote about earlier this year - but it’s not there yet. So either platforms will have to dramatically slow down how fast ads can get up so they can go through human review, or we should expect them to be overly broad in what they include or that a lot get through the cracks.
Ads that are exempt: I’m not wild that ads from official sources aren’t in scope. While they aren’t trying to influence who people vote for, it can influence who gets out to vote and in the spirit of transparency, I think people should know what is being shown to voters. These recommendations also don’t address how to handle ads from news outlets. I firmly believe that the ads a news outlet run and who they choose to target stories to can influence a voter and we should be able to see that. Maryland in fact required this in their law, but a judge said it went too far when news orgs complained saying it puts an undue burden on them.
Sharing info with other service providers: The recommendations strongly encourage and ask that platforms share information amongst one another - including that if an ad is labeled political on one platform the others should do so as well. While I’m all for information sharing, who mediates disagreements of what is a political ad between companies? What is required to determine if an ad is the same given all the various A/B testing that might be done?
Requiring ads to be marked as ads even when shared: For years many people have thought that an ad stays an ad even if it is shared by someone. Others have said an ad is only an ad when someone is seeing it because someone paid for it to be put in front of them versus a friend choosing to do so or if they saw it because they’re a fan of the page. I generally feel this should be a pretty simple solution if platforms would just put some resources to it so that someone can understand when they are seeing an ad vs content that was shared organically but is also being boosted with money. However, whether or not users will understand the distinction is a whole other issue.
Showing people why they are seeing an ad: This is another thing that makes total sense on the surface but executing on is much more difficult. Of course, it would be great if a user could understand every reason why they were served an ad. The fact of the matter is there are so many inputs both from the platform and the advertiser that this will be much harder to map out for people.
Confirming information: It makes total sense that for political and issue ads we know who the advertiser is, how to get in contact with them, etc. However, we can’t just rely on advertisers to tell the truth - so we need ways of verifying this information. This has been a challenge for platforms. I’d like to see in regulations like this what ways platforms should confirm that the information being given to them is correct.
Only need to keep info for one year: For Facebook, we decided to keep ads in the archive for seven years so that researchers, journalists, and others could understand the full scope of ads from one election to the next. Seven was picked because the longest cycle any election has is six years. The Commission’s recommendation is to only keep the ads for one year. This seems way too short to be able to do any adequate research on trends.
If I were a political advertiser in the United States I’d be watching these recommendations from Europe very closely. Should this law pass before anything in the U.S. and the platforms choose to comply rather than just shutting down political and issue ads, then I could see them extending whatever they need to do in Europe to the rest of the world.
What I’m Reading
Washington Post: Facebook reveals broad takedown of global disinformation networks, including some tied to anti-vaccine groups and state actors
The Diplomat: Social Media Misinformation and the 2022 Philippine Elections
Digital Media and Democracy: A Systematic Review of Causal and Correlational Evidence Worldwide
Lego Gutenberg press: Vote for this adorable Lego of the Gutenberg Press so I can get it for myself.
Crypto: I want to get more knowledgeable on all things crypto so I asked the Twittersphere for recommendations. Here’s what they shared: