Today, the European Parliament’s LIBE committee voted to restrict the use of personal data to target online political advertisements to data explicitly provided for this purpose by citizens with their consent, excluding the use of behavioral and inferred intelligence on citizens. LIBE has the exclusive competence on the articles dealing with data protection (the targeting). However, the position will be subject to trilogue negotiations with EU governments.
Specifically, the Committee decided:
- The use of personal data to target online political advertisements would be limited to data explicitly provided for this purpose by citizens with their consent, excluding the use of behavioral and inferred intelligence on citizens (“surveillance advertising”). Refusing consent should be no more complicated than giving it. The “do not track” setting would need to be respected without bothersome prompts. Users who refuse to consent would still have access to online platforms.
- The platforms would be banned from running opaque ad delivery algorithms to determine who should see a political ad; they would only be able to select recipients randomly in the pool of people delineated by the targeting parameters chosen by the sponsor.
- In the 60 days prior to an election or referendum, different political messages may be spread only of the basis of a voter’s language and the constituency they live in, avoiding a fragmentation of the public debate and the sending of contradictory and dishonest messages.
- If a data protection authority such as the Irish DPA fails to enforce the rules against large online platforms, the European Data Protection Board would be able to take over. In cases of illegal political ads targeting it will not only be able to impose financial sanctions but can also temporarily suspend the targeting of ads by advertisers who seriously and systematically violated the rules. This ensures that more affluent sponsors are not able to factor-in the price of financial sanctions in their budget.
- The targeting of political ads on the basis of a person’s racial or ethnic origin, political opinions, religious beliefs, health conditions or sexual orientation is to be banned both offline and online.
- Depending on interpretation, no restrictions might apply to the use of personal data (including behavioural and inferred intelligence as well as sensitive data) by campaigns to target political messages via letter, e-mail, or text messaging at large scale. The targeting rules (Articles 12- and 12) would apply only where external “political advertising services” are used, thus excluding letters, e-mail and text messages sent directly by campaigns.
Pirate MEP Patrick Breyer comments:
“From the Donald Trump and Brexit campaigns we have learned that you can very effectively manipulate a voter if you know which message works on them. While many parties are using personalised targeting, it benefits mostly populist and anti-democratic forces.
I regret that one major loophole remains, however. The loophole for direct messages is bound to be exploited by anti-democratic and anti-european movements to manipulate elections and referendums by spreading disinformation and hatred tailored to each voter’s personality and weaknesses. Their online behaviour and habits right down to sexual orientation or religious beliefs could be exploited to target manipulative and false political messages, as happened in the Trump and Brexit campaigns. We should not allow this Regulation to be circumvented by simply moving from manipulative Facebook Ads to direct messages. We urgently need to close this loophole thus in the upcoming process. For the time being we abstained on the position.”