Last week, Vice President of Global Affairs and Communications Nick Clegg spoke with Richard Waters of the Financial Times, as part of The Next Web conference, about how to balance regulation and innovation.

As part of this discussion, Nick explained the new measures Facebook has put in place to protect the 2020 US presidential election, its efforts to protect elections around the world since 2016, progress in removing harmful content, and the need for a solution to EU-US data flows to avoid economic damage.

The full discussion can be watched above. The following paragraphs are an edited transcript of Nick’s remarks on select topics to provide a snapshot of the points he raised during the conversation.

Since this discussion, we have rolled out several new measures in the run up to the US elections, including prohibit ads that seek to undermine the ultimate outcome of the election and planning to temporarily stop running all social issue, electoral, and political ads in the US after the polls close on November 3, to reduce opportunities for confusion or abuse.

New measures to protect the 2020 US election

This is a polarized and aggressive election at a difficult time for the world and for the country, in the midst of this pandemic and its damaging socioeconomic effects.

One of the things we’re focused on is what happens after the election. We have specific and concrete measures for if a candidate were to declare prematurely victory before the official results are in.

First, we will label the post so that people will see that the results have not yet been certified and that the declaration of victory is premature.

Second, we’re working with Reuters and others to make sure that once the election results are definitively certified, we can convey that at great volume to tens of millions of Americans by prominently displaying them in their newsfeeds.

We are also en route to our target of having 4 million Americans registered to vote and we have a new tool to help us: the Voting Information Center. This will be displayed at the top of people’s News Feeds from now until November 3rd and provides information about how to vote, when to vote, and how to fill in a mail-in ballot according to the local rules.

We have learned from the mistakes of the 2016 election and are far better equipped to protect elections

Facebook has moved with enormous ambition and speed since the mistakes of the 2016 election, which clearly shook the company down to its roots.

What we’ve put in place recently is a final iteration of additional guardrails. These come on top of four years of work to make sure that 2020 is, at least as far as Facebook is concerned, an election which is conducted on our platform better than it was in 2016.

Since then, there have been more than 200 elections around the world, where we have constantly iterated to improve our election defense mechanisms. We’ve employed 35,000 people and we work with 70 fact checkers in over 50 languages. We have an Ads Library to provide transparency on who’s paying for ads, who’s running them, which voters they’re trying to address, and how much money they’re spending. This is far more transparent than anything that exists in the radio, television or print media.

Over the last year, we’ve spent more money — billions of dollars — on the integrity of our platforms than the total revenues of the company when it was floated back in 2012.

We’ve made significant progress detecting and removing harmful content

You are never going to eliminate everything you don’t want online. So the question is, are we able to hold ourselves to account for making significant progress in meeting people’s expectations at scale? Now, on that, we’ve done some very bold new things.

Two years ago we were only able to identify 23% of hate speech before it’s reported to us, but we now do so in over 90% of cases. This shows that our machine learning tools, and in particular AI systems, are able to operate at scale. Last year we removed 6.5 billion fake accounts, which is significant progress.

A new solution for EU-US data flows is needed to avoid a dramatic shock to the data economy

The European Union and the US need to flesh out a solution that enables data to be transferred out of Europe to other jurisdictions. The European Court of Justice first struck down Safe Harbour, then struck down Privacy Shield, and now the Irish Data Protection Commissioner appears to be raising questions about the use of Standard Contract Clauses, which are used by the vast majority of companies in Europe.

At the heart of this is the court’s view that the data protections which are afforded to EU citizens cannot be guaranteed in view of the US surveillance patterns. This is even though those EU data protection provisions explicitly carve out for EU member states surveillance powers which in many ways are just as intrusive as the American surveillance powers.

Like many other companies, we are caught in the middle of this. We’re all looking for legal certainty, which can only be provided by the EU and the US negotiating a successor regime for Privacy Shield. In the meantime, I hope that the ability for companies to transfer data is not suddenly turned off, because that would deliver a dramatic shock to the data economy, in Europe and elsewhere.

We have data centers in Europe and all around the world. The issue is not about where data is located but about who has access to it. The idea that you can divide up global data into perfectly segmented cake slices is not something which anyone who knows anything about the data economy thinks is realistic. But even if you did it, it doesn’t actually answer the underlying issue of access to that data, which is in dispute between the European Union and the US. In the end, the European Union and the US have to resolve this through political negotiation.