Election 2019: What big tech isn't telling us about ads


Social media ads

All over the country, voters, reporters and other observers have been going through something of an awakening about the extent that political parties are using social media to target us.

In a post-Cambridge Analytica-scandal world there’s suspicion about how we might be being manipulated.

To address concerns, the tech giants have created databases to show what political adverts are being run and by whom.

There’s no doubt that more is being disclosed than in previous campaigns, but critics say there’s still much more that “big tech” could reveal.

Let’s start with Facebook.

In October last year, it launched its Ad Library in the UK. Since then, more than 131,400 adverts related to politics, elections, and social issues have been added to the database. That’s almost £11m worth, according to the firm.

Image copyright
Facebook

Image caption

Facebook launched its political ads library in the UK a year ago

The library is free to use and easy to navigate. And recently it began including ads run on the firm’s photo-centric app, Instagram.

You can look up roughly how many times an ad was seen, its approximate cost, the gender and age of those targeted and who made and paid for the advert.

You can also see whether an ad was aimed at people in England, Scotland, Wales and/or Northern Ireland.

Image copyright
Facebook

Image caption

This is the data Facebook provided for one ad run by the Labour Party

However, there’s still much that is not shared.

We know that political parties target voters in very specific areas, such as marginal seats, but the library doesn’t reveal where exactly an ad was shown.

We also know that people are targeted by personal details – for example, interested in “the environment” or “yoga” or more political interests, like “GMB union”. But that information, too, is not shared to the wider public.

Image copyright
Facebook

Image caption

Facebook allows users to be targeted by interests including environmentalism and specific trade unions

Facebook has previously said that transparency is important to “prevent future interference in elections”.

So reporters recently pressed it about these shortcomings. The firm said it planned to improve the database but would not be drawn on whether it intended to divulge the missing information about geographic or interest-based targeting.

When shown an ad, Facebook users can ask for more specific information about how it came to be shown to them.

The BBC and other news organisations have been trying to crowdsource the information from those targeted. But that only gives a fragmented view of what is going on – an unsatisfactory state of affairs, according to the co-founder of the Coalition for Reform of Political Advertising.

“Voters having the full picture of how they’re targeted is important, as it may impact their evaluation of the information being conveyed to them,” Benedict Pringle told the BBC.

“If a voter knows they are being targeted because of their age, sex or occupation it might encourage them to question the message and think, ‘Yes, well you would be saying that to me, wouldn’t you?’

“Also, if techniques deployed by political actors are unavailable for scrutiny by watchdogs and journalists it can enable malpractice.”

But Facebook claims that it is now “more transparent” about ads than those behind more traditional outlets, such as billboards, direct mail, leaflets or targeted email campaigns.

“This is the first UK general election since we introduced these changes and we’re already seeing many journalists using these transparency tools to scrutinise the adverts which are running during this election,” said Rebecca Stimson, head of Facebook’s UK Public Policy.

“This is something we welcome and it’s exactly why we introduced these changes.”

In the 2017 general election, more than £3m was spent by all parties on Facebook.

But a further £1m or so went to another often overlooked source – Google.

Image copyright
Google

Image caption

Google lets users select whether they want to see ads run as text, video or images

Ads can be purchased to appear at the top of its search results, within its YouTube videos or placed on third-party webpages via its ad tech platforms.

So, for example, a user might have been shown brief text-based messages by the Conservative Party promising to make “our streets safer” or the Brexit Party promising to “stand up for democracy”.

Google launched its own version of a political ad library in March, and has so far listed more than 1,400, totalling £144,500.

But it is far more vague than Facebook about the details.

When it lists the amount spent, it does so only in large bands, for example “from £500-£25,000”.

Likewise, the range for the number of people reported to have seen an ad is very broad- for example, “between 10,000 and 100,000”.

Moreover, the site says the details are updated “generally” only every week.

Image copyright
Google

Image caption

Google provides less detail about how ads were targeted than Facebook

There is a way to get more information.

You can find out how ads are targeted by gender, postcode and age. But to do so, you have to download a large, complex set of spreadsheets.

One thing you can’t find out is the crucial “key words” that parties have bought up to ensure their ads are seen when someone searches for certain topics, even those containing rival parties’ names.

Google says it is “thinking hard” about the feedback it’s receiving but has not made any specific commitments.

“We believe that our transparency report helps provide valuable information to voters, public bodies and researchers,” said UK spokesman Elijah Lawal.

He added that it had gone further than Facebook and banned searches based on people’s “interests” in certain sensitive topics, as well as paying regard to the legal protections given against profiling users for having the traits themselves.

“We don’t enable advertisers to target ads to citizens based on their inferred political leanings; nor do we allow ads to be personalised to people based on sensitive information, such as their religion, sexual orientation, or membership in a trade union.”

Snapchat is also running political adverts in this election and it, too, offers detailed targeting to politicians and parties.

The company also has an ad library, albeit only in spreadsheet form.

Earlier this month, Twitter announced it would ban outright political advertising from 22 November. It seems already to have had an effect in the UK, as we’ve not yet seen advertising from any of the main political parties on the platform.

It follows LinkedIn, which introduced a ban in June 2018. TikTok announced a similar measure last month.

For now, all the commitments are voluntary. But some think that it’s time the sector faced official scrutiny.

“Parliament needs to legislate to appoint or create a political-advertising regulator and update laws relating to political advertising,” said Mr Pringle.

“A political-content regulator could then develop a code based on legislation, which would require oversight by Parliament.”

But that won’t happen before this election, and it will be up to the next batch of MPs to decide whether to make it a priority.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *