Online Harms – how does the UK government plan to address election interference?

The UK government has set out a plan to give media regulator Ofcom more powers to regulate internet companies. The talk is about forcing platforms to hold a ‘duty of care’ for their users. We don’t yet know many of the details, and the term online harms covers a vast swathe of activities from child protection to terrorism, but we also know that the government has previously viewed interference with elections as one of the online harms that needs addressing.

So what is an online harm when it comes to electoral integrity and how could internet companies police such threats?

The first is interference with the electoral process itself. In this case, the UK’s adherence to paper and pencil voting and counts taking place in a single venue for each constituency actually helps. If we used electronic voting machines, or voted via the internet, then these might be open to manipulation. The only instance where the process is open to such manipulation is when scanning is used to count the votes in elections such as the mayor of London. So there needs to be confidence that the scanners and their software is secure and accurate and consideration needs to be given to having at least a sample of the paper ballots hand counted by hand.

The next key area is disinformation (or fake news). Should platforms like Facebook ensure that such posts are not altering the course of the election? 

To date the government has been keen to stress that it should be up to voters to decide for themselves what is truthful and what is not when it comes to electoral propoganda. Political adverts are exempt from the ‘legal, decent and honest’ requirements of, say, washing powder adverts. So politicians can say whatever they want on social media or their own websites. (Of course, the government could change their minds and require truthfulness, but this would mean establishing some sort of board to decide on truthfulness and a whole host of other issues.)

Just because the government doesn’t require it doesn’t stop the platforms having their own rules governing political speech. Facebook is the most open, allowing politicians to do and say what they want. They exempt political adverts from fact-checking and have said they are in favour of free speech and allowing voters to decide what is true or not. At the other end of the scale, Twitter has banned political adverts but still allows politicians to say whatever they want in organic tweets. And in the middle comes Google which has restricted the targeting allowed for adverts, but still allows things such as a banner advert which directed users to a site called labourmanifesto.co.uk – which turned out to be a Conservative party advert.

Platform policies are pretty much worldwide. So the UK government’s new initiative will throw down a gauntlet in the shape of a challenge to create UK service conditions reflective purely of UK laws. That has happened in other countries, but the platforms (and US government) have complained bitterly. Such laws have yet to be fully tested in the courts so we wait to see how the platforms will react.

Where the UK government may choose to act would be in the areas of user identification and financial probity. Electoral participation in the UK is limited to UK individuals and companies, and there are limits to the amount that can be spent. So it may be that the government chooses to impose new burdens on internet companies to ensure that only legal contributions can be made and that those responsible for adverts are clearly identifiable. This would take the form of clear ‘imprints’ and an open library to see who has produced what and at what cost.

Finally there is the issue of foreign interference. With participation limited to UK individuals and companies, what action might be proposed to prevent interference from those based overseas – either to seek to advantage a particular candidate or just to create disinformation and confusion?

To this end, crossbench peer Lord Cromwell (*) has tabled a question in the House of Lords:

To ask Her Majesty’s Government, given the “real danger that hostile actors use online disinformation to undermine” the UK’s “democratic values and principles” outlined in their Online Harms White Paper, published in April 2019, what steps they plan to take to empower the proposed independent regulator to require online platforms to take down such material that may be perceived to have an impact on an electoral contest; and what guidance they plan to give to that regulator about how any such decision should be reached.

Of course the government may think election interference should be the responsibility of another body, not Ofcom. Or they may think that there should be no regulation or protection in this area – although that would contradict their main decision and they would have to explain why. We await the further details.

 

* Disclosure – Lord Cromwell manages many UK observer secondments to OSCE/ODIHR international election observation missions on behalf of the UK Foreign and Commonwealth Office and has employed me in this role.

Social media and elections – an update

I’ve written a fair amount over the past year about the problems social media companies face with regard to elections and political posts. You can find examples here, here and here.

In short, the argument is that these companies are massively powerful and can have a huge influence on elections. They have chosen different paths with Twitter banning political adverts, Google restricting the targeting that is allowed for such adverts and Facebook regarding the matter as one of free speech and therefore not really making any changes. Plus there is the problem that all companies are working on the basis of a single worldwide policy which pays no regard to the individual laws that might be applicable in the different countries they operate in.

There have been a few updates in recent weeks:

  • Wired has an article which explains the new Facebook Board, a group of people operating at arms length from the platform who will take the final decision on content moderation. As the article points out, they are almost certain to hear a case on false statements in political adverts and, whilst their rulings don’t automatically set a precedent, it may well be that this is the start of a process that leads to a change in policy.
  • Facebook has also updated its political ad library to make it more transparent and given users the option of seeking fewer political adverts in the future.
  • This decision is not without its critics. In the US, many political consultants – used to being able to rely on Facebook’s micro-targeting functions – have suggested that the ability of users to limit the political adverts they see could make campaigning harder. They have produced a report, which it seems Facebook is looking closely at, suggesting ways forward. However, as with so many concerns in the past, this is a purely US campaign industry solution which doesn’t take account of worldwide issues.
  • Twitter came under fire following a BBC exposé which showed that adverts could be targeted at extreme groups such as neo-Nazis. The platform has pledged to ban such adverts in the future.

Are Facebook finally making a move to limit micro-targeting?

The Wall Street Journal says that Facebook is considering taking steps to limit micro-targeting – the practice of allowing advertisers to send individual ads to just a hundred or so users.

If such a move happens then it will be a response to the pressure the platform is feeling from politicians and activists across the world and the moves made by rivals Twitter and Google in recent weeks.

The proposal, according to the WSJ, would be to raise the minimum number of user targets to a few thousand. That would still allow a high degree of granulation – the ability to target ad recipients based on refined characteristics such as personal likes or geographic location. It isn’t the same as Google’s proposal to limit targeting to age, gender and postcode.

If it happens, this is once again a small responsive step from the biggest platform. What we are still missing is the big picture – where do they see their advertising policy being in five years time and how do they respond to the calls from around the world to make political adverts more transparent and, well, truthful. It would be great if Facebook would set out this vision for us rather than scattershot, incremental steps.

Incidentally, this video of comedian Sacha Baron Cohen ripping into the big platforms whilst receiving an award from the US Anti Defamation League is well worth watching in full.

Google cuts back on political adverts

Google will no longer allow adverts featuring deep fakes, micro-targetting or targetting voters based on registered political affiliation according to a letter sent by the company in the US. It will also expand its library of political ads. The decision is being seen as putting further pressure on Facebook.

The Google announcement, in common with other platforms,introduces terms and restrictions based around the current US market and political system. Adverts based on voter affiliation are already banned in the UK and questioning of the census – another restriction – is a US phenomenon. However the measures will apply worldwide and will be put in place before the UK votes on 12th December.

At present, both Labour and Conservatives are investing heavily in Google Ads which appear when users search for the names of other parties. The new policies would not appear to limit such adverts.

“Whether you’re running for office or selling office furniture, we apply the same ads policies to everyone; there are no carve-outs,” said Google Ads executive Scott Spencer in a blogpost. This is widely interpreted as a dig at Facebook which has exempted political adverts from fact-checking.

However, the company has admitted that its resources to check adverts are limited and the number that may be banned as a result of this policy will be small.

Micro-targetting is Facebook’s key advantage and only exists to a very limited extent on Google. Facebook has a vast database of knowledge on each of its users and sells this knowledge to advertisers, including political advertisers. Google will allow adverts based on gender, age and locations as small as individual postcodes – but will not permit other data to be used.

Twitter’s new policy has now been given in more detail and will ban all adverts from politicians and parties and those which are based on specific, even if they come from pressure groups or individuals. The company also aims to ban advertisements aimed at influencing legislation, but not those which refer to generic issues. Campaign groups have claimed that this means that polluting fossil fuel companies can still run ads to promote their products but that campaigners aiming to stop them will be banned.

Google, like Twitter, does not rely heavily on political adverts for revenue, generating around $128m from the US market since June 2018. But the Guardian has revealed that the company has been under-reporting poltical spending in the UK by a large factor. The newspaper claims that Labour was reported as having spent £50 in the week beginning 27th October, but that the actual figure was around £63,000. A smaller discrepancy also existed for the Conservatives. Whilst there is no indication that Labour or the Conservatives intended to mislead regulators, accurate reporting by platforms is vital to enable the Electoral Commission and public to check that parties are not over-spending on the election.