A Luxembourg survey published by Atoz Tax advisors in 2023 distilled the ideal politician as “reliable, open and communicative.” What better way to give the appearance of openness and reliability than communicating online?
The average person spends almost seven hours online every day. Compared to the seconds spent passing a street billboard, the argument for political campaigning online needs little introduction. However, it’s not just because voters are ever present online that parties choose to campaign there.
The core appeal of online campaigning is the speed and anonymity with which a message can be deployed, combined with the portability of the space. If something happens that a party could potentially leverage, they can respond within minutes, if not seconds.
Costs can also be significantly lower than traditional approaches. It is possibly one of the reasons why in the 2018 legislative elections the Pirate Party placed ads on adult entertainment website Pornhub.
High quality video filming and editing features in smartphones and on social media platforms also lower the barriers for generating organic content like videos, something which a large number of individual candidates in Luxembourg took advantage of ahead of the 2023 legislative elections.
And today, online tools enable parties to better target their messages to voters. When posting an advert on Meta, it has never been easier to set the parameters to target a specific demographic.
So, what’s new?
What is new is the trend of microtargeting: “Turning to individual voter’s authentic preferences, these can be microtargeted with individually tailored advertising based on their digital footprints–a process that feeds back powerfully into preference formation. As with offline campaigns, simplifications and distortions may often be in the mix,” writes the International Idea Institute for Democracy and Electoral Assistance in its report “Regulating Online Campaign Finance”.
Data-driven politics
The concept isn’t purely pernicious. Microtargeting can also be used to hone a party’s policies around voter needs, argues Rachel Gibson, a leading expert in party politics and digital campaigns. In an October 2023 interview with the European Research Council, she pointed out that the data gathered in the process of microtargeting around individual voter preferences, personalities and psychological profiles “could be used to create nuanced profiles to serve as the foundation for making strategic decisions regarding your political campaign. It informs everything from the content of your messages and the intended recipients to the choice of communication channels.”
Gibson says that parties have yet to transition to a completely data-driven approach, although early signs suggest a more prevalent shift in this direction in the US.
Another more recent trend has seen political parties leverage influencers and digital content creators to amplify messages. In the 2020 elections, Joe Biden’s campaign hired a team of 25 to work on its influencer campaign in a bid to reach younger voters. Incidentally, he won voters under the age of 29 by a 26-point margin.
Bots & Trolls
An approach which has exploded in the last decade is the use of automated software programmes, or bots, to accelerate the spread of or denigrate a message on social media.
In 2017, a group of activists built a bot that would “automate flirty exchanges with real people in the UK, conversations which eventually switched to the strengths of the UK Labour Party. The bot sent somewhere between 30,000 and 40,000 messages, targeting 18- to 25-year-olds in constituencies where the Labour candidates were running in tight races.
An emerging white hat trend within this domain is the use of chatbots. In Luxembourg, the Demokratesch Partei (DP) integrated an interactive video chatbot into the party’s website to present its electoral themes to voters in the run-up to the 2023 legislative elections. Speakers in the short videos included then prime minister Xavier Bettel, finance minister Yuriko Backes, SME minister Lex Dells and civil service minister Marc Hansen, with subtitles in English, French and German.
“For politicians, this presents a novel way of engaging new audiences online, including younger potential voters like Gen Z, who are digitally native and prefer to consume online media in video format,” Videobot, which helped develop the chatbot, said.
Bad bots
According to tech publication IEEE Spectrum, bots, along with trolls, have also been associated with inauthentic behaviour as far back as 2010 to provoke, to harass, intimidate, or confuse, often on the payroll of non-domestic political organisations. For example, investigative journalism group Correctiv recently exposed a pro-Russian fake account network in Germany spreading misleading pro-Kremlin narratives, links to disinformation sites and fake government documents via Meta adverts. It is believed the troll activity was intended to influence the outcome of the 2024 European parliamentary election.
Cutting through the fakes
Since the release of ChatGPT in November 2022, generative AI tools capable of mimicking reality have become increasingly accessible. This opens a Pandora’s box of opportunities for nefarious political campaigns to proliferate through fake news, videos, audio and photos.
In August 2023, Poland’s largest opposition party, the centrist Civic Platform (PO), broadcast an election campaign advert combining genuine clips of words spoken by the prime minister with fake ones generated by AI. In September 2023, in the US, a video of Florida Governor Ron DeSantis announcing that he was dropping out of the 2024 presidential race circulating on social media was found to be a deep fake video created by artificial intelligence.
Rachel Gibson says we have yet to witness a major scandal regarding a fabricated AI generated viral story that changes the way people vote. But that day will come.
It doesn’t help that social networks are unable or unwilling to police sophisticated AI content. And if they do introduce measures such as labelling AI-generated content, to do so after content has been seen by millions may not be enough to reverse the damage the fake has done.
Just as worrying is the ability of dishonest politicians to weaponise scepticism around AI-generated content. They could easily brush off allegations made against them in authentic video or audio as deep fakes deliberately created by opponents, a strategy known as the “liar’s dividend”.
Lack of oversight
One of the biggest challenges for democracies is the lack of transparency, monitoring, and regulation related to how politicians use online campaigning. Online campaign expenditure is not straightforward to audit since each social media platform has its own definition of political advertising and sets its own policies.
“For instance, Facebook’s (including Instagram) and Snapchat’s definition of political ads include advertisements on social issues, while Google does not treat ads about social issues as political advertisement, while streaming services, such as Hulu and Pandora, do not define what constitutes a political ad,” The International Idea Institute for Democracy and Electoral Assistance explains.
There are similar blurred lines when it comes to what degree users can micro target ads. And how does one treat organic campaign content such as text or video posts made by candidates on their personal feeds?
Even harder to regulate are content and services solicited on the dark web where, for the right price, a candidate could create and manage a troll army or purchase voter datasets. The absence of regulations “provides incentives for political parties to evade reporting of expenditures. This, in turn, makes it difficult for oversight agencies to trace payments and verify the accuracy of financial reporting by the parties concerned,” the Idea Institute writes.
European solution
Europe offers some hope. In November 2023, the European Parliament and the Council reached a decision on the Regulation on transparency of political advertising stating that political adverts must be clearly labelled and must indicate who paid for them, how much, to which elections, referendum or regulatory process they are linked and whether they have been targeted. Targeting and amplification techniques will only be available for online political advertising based on personal data collected from the data subject and subject to consent. The use of sensitive personal data will be banned to avoid the potential for voters to be manipulated. Meanwhile, sponsored ads from outside the EU will be prohibited three months before elections. While a promising step, it will take time for member states to adapt national legislation in-line with the regulation. At the current speed of innovation, by then there may be whole new challenges.
This article was first published in the Silicon Luxembourg magazine. Read the full digital version of the magazine on our website, here. You can also choose to receive a hard copy at the office or at home. Subscribe now.
