Advertisement

Facebook identifies 'political influencers' to spread ads to their friends online

Facebook has developed technology to target “political influencers” to spread paid-for messages to their friends and followers: Getty
Facebook has developed technology to target “political influencers” to spread paid-for messages to their friends and followers: Getty

Facebook has developed technology to target “political influencers” to spread paid-for messages to their friends and followers.

The social media giant, already at the centre of a data misuse scandal, has patented a system using a type of artificial intelligence to identify “politically engaged” and “respected” users who could help push campaign adverts for parties or other organisations.

Facebook insists users would have to opt in to the system through their privacy settings, and new European GDPR data rules state they must give “explicit consent”.

But it is unclear if the recipients would be aware that they had been targeted by these political influencers.

Facebook has previously said it seeks political influencers for better campaign “reach”, but until now it has revealed few details about the technology behind the strategy.

Details have now emerged in a patent application, titled Systems and Methods for Identifying Politically Influential Users, that was filed in the US last year and published in Europe last week.

Key to the powerful software is the ability to identify how politically engaged an individual is, how much they share messages and the number of views and the level of engagement with their content.

The patent describes how the system identifies “credible or respected” sources on politics through natural language processing, which is combined with the number of “engagements” on their posts, such as shares, likes and comments.

The political influencer can be classed in a particular ideology group, possibly based on the individual’s connections, meaning adverts can be targeted at them in the expectation that they will be shared.

They can even be put into sub-categories such as “very liberal”, “conservative” or “moderate politically”.

Deploying machine learning AI therefore helps advertisers — including “political entities, such as politicians, political campaigns, and/or their supporters” — to ensure their messages are more likely to reach a wider audience, according to the patent document.

The system was built by a computational sociologist, who designed computer modules marked “political influence”, “political ideology” and “advertisement” that are connected to Facebook’s sprawling databases.

Greg Epps, co-founder of React Robotics, said deploying cutting-edge AI allows Facebook to pinpoint users’ beliefs on a huge scale.

Damian Collins, Conservative chairman of the Commons digital, culture, media and sport committee, said: “People should have the right to know whether what they are seeing is an advert and who is sending it to them, even if it’s through a third party.”

Facebook promised to be more transparent after the Cambridge Analytica data-harvesting scandal, where personal details of up to 87 million users were allegedly “improperly shared”.

A Facebook spokesman said: “We often seek patents for technology we never implement, and patents should not be taken as an indication of future plans.”