The EU is calling on eight major tech companies including Google, Facebook and X to detail how they identify and tackle deepfake material amid concerns about the use of the technology to influence elections.
In a world first, they will be using new laws on artificial intelligence to force companies to root out fake video, imagery and audio.
Companies have until 5 April to show how they will deal with last-minute, high-impact fake news being dumped on social media, amid evidence that foreign agents, including those from Russia, are building up “sleeper” accounts to be deployed on the eve of elections.
The EU is also launching a formal investigation into the Chinese online trader AliExpress, which has 104 million users in the bloc, over allegations that its marketplace was used to trade online in illegal products such as fake medicines and fake food, and failed to prevent children accessing pornographic material.
The AliExpress investigation is the third official action taken under the Digital Services Act (DSA), one of a pair of online laws passed by the bloc last year alongside the Digital Markets Act (DMA). The EU has previously issued Twitter (now X) and TikTok with formal investigations under the law.
On Thursday, it also announced it was investigating LinkedIn after reports from civil society groups accused the jobs social network of breaches over online advertising profiling.
The eight companies that have been asked to provide more information on their ability to spot faked artificial intelligence (AI) content are Bing, Google Search, Facebook, Instagram, Snapchat, TikTok, YouTube, and X.
Officials noted that not all AI content, which must be labelled as fabricated material under the new EU AI Act, was harmful but that they needed to signal to all tech companies that the new laws were now in place.
Although there have been few major instances of such technology being used for political ends, X, came under particular fire for its failure to prevent the proliferation of a series of pornographic deepfakes of singer Taylor Swift earlier this year.
The commission is calling in all the social media companies to show what they have in place to combat deepfakes. Such risks are not hypothetical.
A report by the EU’s external action service in January showed how Russia had built up a network of fake accounts over a period of months to make them look legitimate just before last year’s general election in Spain.
They were then activated on the eve of the ballot to share generated misinformation, including claims that there were bombs at polling stations.
“We want to be prepared as best as possible. We want to push the platforms to tell us what they are doing to be as well prepared as possible for 11th-hour interventions distributed in preprepared channels at large scale,” an official said.
The EU said it would be creating an “enforcement ecosystem” along with regulators in 27 member states to send messages out to all the platforms that fake material is now illegal under the Digital Services Act.
Under the DSA and the new AI Act companies will be required to have systems in place to “escalate” complaints about fake material.
But officials said they were also keen to send a signal to smaller players and bedroom creators who can spread fake material at speed and at low cost through the large platforms.
Initially announced in November 2023, the investigation into AliExpress is specifically over the measures it has taken to comply with obligations related to risk assessments and mitigation measures to protect consumers online.
The DSA’s sister law, the DMA, has also risen in prominence this month. The DMA’s rules regarding “gatekeepers” – firms with significant power to warp digital markets – came into effect last week, requiring companies including Apple, Google and Meta to change how some of their largest platforms operate.
Smartphone users in the EU will now be offered the ability to bypass the app stores and payment systems run by Apple and Google, while users of WhatsApp will shortly be able to send messages using the platform to other, compatible, apps.
AliExpress said: “We respect all applicable rules and regulations in the markets where we operate. As a VLOP [very large online platform], we have been working with, and will continue to work with, the relevant authorities on making sure we comply with applicable standards and will continue to ensure that we will be able to meet the requirements of the DSA. AliExpress is committed to creating a safe and compliant marketplace for all consumers.”
LinkedIn was approached for comment.
